A longitudinal study of topic classification on Twitter

View article
PeerJ Computer Science

Introduction

With the emergence of the Social Web in the mid-2000s, the Web has evolved from a static Web, where users were only able to consume information, to a Web where users are also able to interact and produce information (Bouadjenek, Hacid & Bouzeghoub, 2016). This evolution, which is commonly known as the Social Web, has introduced new freedoms for the user in their relation with the Web by facilitating their interactions with other users who have similar tastes or share similar resources. Specifically, social media platforms such as Twitter are commonly used as a means to communicate with other users and to post messages that express opinions and topics of interest. In 2019, it was estimated that more than 330 million users posted 500 million tweets per day (https://www.brandwatch.com/blog/twitter-stats-and-statistics/).

Consequently, Twitter represents a double-edged sword for users. On one hand it contains a vast amount of novel and topical content that challenge traditional news media sources in terms of their timeliness and diversity. Yet on the other hand Twitter also contains a vast amount of chatter and otherwise low-value content for most users’ information needs where manual filtering of irrelevant content can be extremely time-consuming. Previous work by Lin, Snow & Morgan (2011), Yang et al. (2014) and Magdy & Elsayed (2014) has noted the need for topic-based filtering on Twitter and has proposed a range of variations on supervised classification techniques to build effective topic filters.

While these previous approaches have augmented their respective topical classifiers with extensions including semi-supervised training of multiple stages of classification-based filtering and online tracking of foreground and background language model evolution, we seek to analyze the lowest common denominator of all of these methods, namely the performance of the underlying (vanilla) supervised classification paradigm. Our fundamental research questions in this article are hence focused on a longitudinal study of the performance of such supervised topic classifiers. For example, over a year or more after training, how well do such classifiers generalize to future novel topical content, and are such results stable across a range of topics? In addition, how robust is a topic classifier over the time horizon, e.g., can a model trained in 1 year be used for making predictions in the subsequent year? Furthermore, what features, feature classes, and feature attributes are most critical for long-term classifier performance?

To answer these questions, we collected a corpus of over 800 million English Tweets via the Twitter streaming API during 2013 and 2014 and learned topic classifiers for 10 diverse themes ranging from social issues to celebrity deaths to the “Iran nuclear deal”. We leverage ideas from Lin, Snow & Morgan (2011) for curating hashtags to define our 10 training topics and label tweets for supervised training; however, we also curate disjoint hashtag sets for validation and test data to tune hyperparameters and evaluate true generalization performance of the topic filters to future novel content.

The main outcomes of this work can be summarized as follows:

  • We empirically show that the random forest classifier generalizes well to unseen future topical content (including content with no hashtags) in terms of its average precision (AP) and Precision@n (for a range of n) evaluated over long time-spans of typically 1 year or more after training.

  • We demonstrate that the performance of classifiers tends to drop over time–roughly 35% drop in Mean Average Precision 350 days after training ends, which is an expected, but nonetheless significant decrease. We attribute this to the fact that over long periods of time, features that are predictive during the training period may prove ephemeral and fail to generalize to prediction at future times.

  • To address the problem above, we show that one can remove tweets containing training hashtags from the validation set to allow better parameter tuning leading to less overfitting and improved long-term generalization. Indeed, although our approach here is simple, it yields a roughly 11% improvement for Mean Average Precision.

  • Finally, we provide a detailed analysis of features and feature classes and how they contribute to classifier performance. Among numerous insights, we show that the class of hashtags and simple terms have some of the most informative feature instances. We also show that the volume of tweets by a user correlates more with their informativeness than their follower or friend count.

In summary, this work1 provides a longitudinal study of Twitter topic classifiers that further justifies supervised approaches used in existing work while providing detailed insight into feature properties and training methodologies leading to good performance. The rest of this article is organized as follows: we first review the literature and then describe the notation we use in this article as well as a formal definition of the problem we address. Next, we provide a description of the dataset we used for the analysis in this article, followed by a description of the general methodology we use for learning topic classifiers. Finally, we provide a discussion of our empirical results before concluding and outlining future work.

Related work

There is a substantial body of research related to topic classification in social media. Below, we review the major works related to Twitter topic classification, topic modeling for social media and applications of classifiers for social media (including tweet recommendation, event detection in social media, and “friend sensors”).

Twitter topic classification

Topic classification for social media aims to detect and track general topics such as “Baseball” or “Fashion”. In previous work, researchers have collected labeled data either by using a single hashtag for each topic (Lin, Snow & Morgan, 2011), a user-defined query for each topic (Magdy & Elsayed, 2014), manual labeling (Daouadi, Zghal Rebaï & Amous, 2021; Ayo et al., 2021), or co-training based on the URLs and text of the tweet (Yang et al., 2014). We expand on Lin, Snow & Morgan (2011)’s work and use a set of hashtags instead of a single hashtag. Similarly, we extract features consisting of hashtags, mentions, unigram terms, and authors as done in this prior work, but also add location as another feature, which has shown to be the second most important feature for topic classification after unigram terms. Furthermore, we provided a novel learning and evaluation paradigm based on splitting both the data and hashtags along temporal boundaries to generate train, validation and test datasets in order to evaluate long-term generalization of trained topic classifiers. In contrast, we remark that Lin, Snow & Morgan (2011) only evaluated over 1 week, Magdy & Elsayed (2014) over 4 days, and Yang et al. (2014) did not explicitly mention the data duration or that their study was intended to assess long-term performance. Hence these previous studies do not permit one to assess the long-term topic classification performance of topic classifiers for Twitter as intended by the 2 year longitudinal study performed in this article.

Topic modeling for social media

Topic models are a type of statistical model for discovering abstract “topics” that occur in a collection of documents (Blei, 2012). For this purpose, machine learning researchers have developed a suite of algorithms including Probabilistic Latent Semantic Analysis (PLSA) (Hofmann, 1999), Non-negative matrix factorization (Lee & Seung, 1999; Arora, Ge & Moitra, 2012; Luo et al., 2017), and Latent Dirichlet allocation (LDA) (Blei, Ng & Jordan, 2003). LDA is perhaps the most common topic model currently in use.

While topic models such as LDA have a long history of successful application to content domains such as news articles (Chen et al., 2010; Cohen & Ruths, 2013; Greene & Cross, 2015) and medical science (Paul & Dredze, 2011; Wu et al., 2012; Zhang et al., 2017), they are often less coherent when applied to social media and specifically microblog content like Twitter. In particular, Twitter poses challenges for topic modeling mainly because it contains short and messy text (Zhao et al., 2011b; Han, Cook & Baldwin, 2012; Mehrotra et al., 2013; Jelodar et al., 2018; Zuo et al., 2021). This problem has been frequently addressed through content pooling methods (Hong & Davison, 2010; Weng et al., 2010; Naveed et al., 2011; Mehrotra et al., 2013; Alvarez-Melis & Saveski, 2016), which comprise a data preprocessing step consisting of merging related tweets together and presenting them as a single document to the topic modeling algorithm. In a different vein, several works proposed to integrate network structure with topic modeling (Tang et al., 2008; Chen, Zhou & Carin, 2012b; Kim et al., 2012; Chen et al., 2017). Very recent work by Nolasco and Oliveira (Nolasco & Oliveira, 2019) proposed a method for detecting subevents within main complex events through topic modeling in social media posts.

Despite this rich tradition of work in topic modeling including applications to Twitter, we remark that all of these methods are unsupervised and seek to discover topics, whereas our work is focused on the supervised setting where topics (and their labels) are available and we are concerned with long-term classifier accuracy in this supervised, known topic setting.

Related applications of classifiers for social media

Aside from highly related work on supervised topic classifiers for Twitter (Lin, Snow & Morgan, 2011; Yang et al., 2014; Magdy & Elsayed, 2014) that motivated this study as discussed previously, there are many other uses of classifiers for social media. While we argue no prior work has performed a longitudinal analysis of supervised Twitter topical classifiers as done in this article, these alternative applications of classifiers for social media may broadly benefit from the insights gained by our present study. We cover these related uses below along with important differences with the present work, divided into the following four subareas: (1) trending topic detection, (2) tweet recommendation, (3) friend sensors, and (4) specific event detection such as earthquake or influenza sensors.

Trending Topic Detection represents one of the most popular types of topical tweet detector and can be subdivided into many categories. The first general category of methods define trends as topically coherent content and focus on clustering across lexical, linguistic, temporal and/or spatial dimensions (Petrović, Osborne & Lavrenko, 2010; Ishikawa et al., 2012; Phuvipadawat & Murata, 2010; Becker, Naaman & Gravano, 2011; O’Connor, Krieger & Ahn, 2010; Weng & Lee, 2011). The second general category of methods define trends as temporally coherent patterns of terms or keywords and focus largely on detecting bursts of terms or phrases (Mathioudakis & Koudas, 2010; Cui et al., 2012; Zhao et al., 2011a; Nichols, Mahmud & Drews, 2012; Aiello et al., 2013). The third category of methods extends the previous categories by additionally exploiting network structure properties (Budak, Agrawal & El Abbadi, 2011). Despite this important and very active area of work that can be considered a type of topical tweet detector, trending topic detection is intrinsically unsupervised and not intended to detect targeted topics. In contrast, the work in this article is based on supervised learning of a specific topical tweet detector trained on the topical set of hashtags provided by the user.

Tweet Recommendation represents an alternate use of tweet classification and falls into two broad categories: personalized or content-oriented recommendation and retweet recommendation. For the first category, the objective of personalized recommendation is to observe a user’s interests and behavior from their user profile, sharing or retweet preferences, and social relations to generate tweets the user may like (Yan, Lapata & Li, 2012; Chen et al., 2012a). The objective of content-oriented recommendation is to use source content (e.g., a news article) to identify and recommend relevant tweets (e.g., to allow someone to track discussion of a news article) (Krestel et al., 2015). For the second category, there has been a variety of work on retweet prediction that leverages retweet history in combination with tweet-based, author-based, and social network features to predict whether a user will retweet a given tweet (Can, Oktay & Manmatha, 2013; Xu & Yang, 2012; Petrovic, Osborne & Lavrenko, 2011; Gilabert & Segu, 2021). Despite the fact that all of these methods recommend tweets, they—and recommendation methods in general—are not focused on a specific topic but rather on predicting tweets that correlate with the preferences of a specific user or that are directly related to specific content. Rather the focus with learning topical classifiers is to learn to predict for a broad theme (independent of a user’s profile) in a way that generalizes beyond existing labeled topical content to novel future topical content.

Specific Event Detection builds topical tweet detectors as we do in this work but focuses on highly specific events such as disasters or epidemics. For the use case of earthquake detection, an SVM can be trained to detect earthquake events and coupled with a Kalman filter for localization (Sakaki, Okazaki & Matsuo, 2013), whereas in Bouadjenek, Sanner & Du (2020), Bouadjenek & Sanner (2019) a relevance-driven clustering algorithm to detect natural disasters has been proposed. In another example use case to detect health epidemics such as influenza, researchers build purpose-specific classifiers targeted to this specific epidemic (Culotta, 2010; Aramaki, Maskawa & Morita, 2011), e.g., by exploiting knowledge of users’ proximity and friendship along with the contageous nature of influenza (Sadilek, Kautz & Silenzio, 2012). While these targeted event detectors have the potential of providing high precision event detection, they are highly specific to the target event and do not easily generalize to learn arbitrary topic-based classifiers for Twitter as analyzed in this work.

Friend Sensors are a fourth and final class of social sensors intended for early event detection (Kryvasheyeu et al., 2015; Garcia-Herranz et al., 2014) by leveraging the concept of the “friendship paradox” (Feld, 1991), to build user-centric social sensors. We note that our topical classifiers represent a superset of friend sensors since our work includes author features that the predictor may learn to use if this proves effective for prediction. However, as shown in our feature analysis, user-based features are among the least informative feature types for our topical classifier suggesting that general topical classifiers can benefit from a wide variety of features well beyond those of author features alone.

Notation and problem definition

Our objective in this article is to carry out a longitudinal study of topic classifiers for Twitter. For each Twitter topic, we seek to build a binary classifier that can label a previously unseen tweet as topical (or not). To achieve this, we train and evaluate the classifier on a set of topically labeled historical tweets as described later in this article.

Formally, given an arbitrary tweet d (a document in text classification parlance) and a set of topics T = {t1,…,tK}, we wish to train ft(d) to predict a continuous score value for each topic tT over a subset of labeled training tweets from D = {d1,…,dN }. We assume that each tweet diD (for i ∈ {1,…,N }) is represented by a vector of M binary features di=[di1,,diM] with dim{0,1} (for m ∈ {1,…,M }) indicating that the mth feature occurs in di (1) or not (0). Each tweet di also has an associated topic label t(di){0,1} to indicate whether the tweet di is topical (1) or not (0). As done in many standard classifiers (e.g., naïve Bayes, logistic regression, SVM), we wish to learn a scoring function ft(d) such that a higher score ft(d) indicates a higher confidence that d should classified as topical for t and furthermore this generalizes well to new unseen tweet data not encountered during training.

Data description

We begin with details of the Twitter testbed for topical classifier learning that we evaluate in this article. We crawled Twitter data using Twitter Streaming API for 2 years spanning 2013 and 2014 years. We collected more than 2.5 TB of compressed data, which contains a total number of 811,683,028 English tweets. In the context of Twitter, we consider five feature types for each tweet. Each tweet has a User feature (i.e., the person who tweeted it), a possible Location (i.e., a string provided as meta-data), and a time stamp when it was posted. A tweet can also contain one or more of the following:

  • Hashtag: a topical keyword specified using the # sign.

  • Mention: a Twitter username reference using the @ sign.

  • Term: any non-hashtag and non-mention unigrams.

We provide more detailed statistics about each feature in Table 1. For example, there are over 11 million unique hashtags, the most frequent unique hashtag occurred in over 1.6 million tweets, a hashtag has been used on average by 10.08 unique users, and authors (Users) have used a median value of 2 tweets.

Table 1:
Feature Statistics of our 811,683,028 tweet corpus.
#Unique features
User Hashtag Mention Location Term
85,794,831 13,607,023 46,391,269 18,244,772 16,212,640
Feature usage in #Tweets
Feature Max Avg Median Most frequent
User 10,196 8.67 2 running_status
Hashtag 1,653,159 13.91 1 #retweet
Mention 6,291 1.26 1 tweet_all_time
Location 10,848,224 9,562.34 130 london
Term 241,896,559 492.37 1 rt
Feature usage by #Users
Hashtag 592,363 10.08 1 #retweet
Mention 26,293 5.44 1 dimensionist
Location 739,120 641.5 2 london
Term 1,799,385 6,616.65 1 rt
Feature using #Hashtags
User 18,167 2 0 daily_astrodata
Location 2,440,969 1,837.79 21 uk
DOI: 10.7717/peerj-cs.991/table-1

Figure 1 shows per capita tweet frequency across different international and U.S. locations for different topics. While English speaking countries dominate English tweets, we see that the Middle East and Malaysia additionally stand out for the topic of Human Caused Disaster (MH370 incident), Iran, U.S., and Europe for nuclear negotiations the “Iran Nuclear deal”, and soccer for some (English-speaking) countries where it is popular. For U.S. states, we see that Colorado stands out for health epidemics (both whooping cough and pneumonic plague), Missouri stands out for social issues (#blacklivesmatter in St. Louis), and Texas stands out for space due to NASA’s presence there.

(A–F) Per capita tweet frequency across different international and U.S. locations for different topics.

Figure 1: (A–F) Per capita tweet frequency across different international and U.S. locations for different topics.

The legend provides the number of tweets per 1 million capita.

Methodology

In this section, we describe the formal framework we use for our longitudinal study of topic classification. We begin by describing how we automatically label data using a set of manually curated hashtags. Then, we proceed to describe how we temporally split both the dataset and manually curated hashtags into train, validation and test sets, which is a critical step for our longitudinal study of topical classifiers and long-term generalization. Finally, we provide a brief description of several score-based classification algorithms and one ranking algorithm used in our analysis.

Dataset labelling

A critical bottleneck for learning targeted topical social classifiers is to achieve sufficient supervised content labeling. With data requirements often in the thousands of labels to ensure effective learning and generalization over a large candidate feature space (as found in social media), manual labeling is simply too time-consuming for many users, while crowdsourced labels are both costly and prone to misinterpretation of users’ information needs. Fortuitously, hashtags have emerged in recent years as a pervasive topical proxy on social media sites—hashtags originated on Internet Relay Chat (IRC), were adopted later (and perhaps most famously) on Twitter, and now appear on other social media platforms such as Instagram, Tumblr, and Facebook. Following the approach of Lin, Snow & Morgan (2011), for each topic tT, we leverage a (small) set of user hand-curated topical hashtags Ht to efficiently label a large number of supervised topic labels for social media content.

Specifically, we manually curated a broad thematic range of 10 topics shown in the top row of Table 2 by annotating hashtag sets Ht for each topic tT. We used four independent annotators to query the Twitter search API to identify candidate hashtags for each topic, requiring an inter-annotator agreement of three annotators to permit a hashtag to be assigned to a topic set. Samples of hashtags for each topic are given in the bottom row of Table 2.

Table 2:
Train/Validation/Test Hashtag samples and statistics.
Tennis Space Soccer Iran nuclear deal Human disaster Celebrity death Social issues Natural disaster Epidemics LGBT
#TrainHashtags 62 112 144 12 57 33 37 61 55 30
#ValHashtags 14 32 42 2 8 4 5 4 17 9
#TestHashtags 14 17 21 3 12 7 8 17 13 5
#+TrainTweets 21,716 5,333 14,006 6,077 153,612 155,121 27,423 46,432 14,177 1,344
#-TrainTweets 191,905 46,587 123,073 54,045 1,363,260 1,376,872 244,106 411,609 125,092 11,915
#+ValTweets 884 2,281 4,073 1,261 53,340 23,710 3,088 843 4,348 50
#-ValTweets 7,860 20,368 36,341 11,363 473,791 210,484 27,598 7,456 39,042 443
#+TestTweets 1,510 5,908 11,503 368 34,055 7,334 14,566 5,240 3,105 692
#-TestTweets 13,746 53,348 103,496 3,256 305,662 65,615 130,118 47,208 27,828 6,325
Sample Hashtags #usopenchampion #asteroids #worldcup #irandeal #gazaunderattack #robinwilliams #policebrutality #earthquake #ebola #loveislove
#novakdjokovic #astronauts #lovesoccer #iranfreedom #childrenofsyria #ripmandela #michaelbrown #storm #virus #gaypride
#wimbledon #satellite #fifa #irantalk #iraqwar #ripjoanrivers #justice4all #tsunami #vaccine #uniteblue
#womenstennis #spacecraft #realmadrid #rouhani #bombthreat #mandela #freetheweed #abfloods #chickenpox #homo
#tennisnews #telescope #beckham #nuclearpower #isis #paulwalker #newnjgunlaw #hurricanekatrina #theplague #gaymarriage
DOI: 10.7717/peerj-cs.991/table-2

Dataset splitting

In the following, we describe key aspects related to the temporal splitting of the dataset and Ht labels for training, validation parameter tuning, and test evaluation purposes. We also outline a methodology for sampling negative examples and an overall training procedure including hyperparameter tuning.

Temporal splits of data and Ht for training, validation and testing: As standard for machine learning methods, we divide our training data into train, validation, and test sets—the validation set is used for hyperparameter tuning to control overfitting and ensure generalization to unseen data. As a critical insight for topical generalization where we view correct classification of tweets with previously unseen topical hashtags as a proxy for topical generalization, we do not simply split our data temporally into train and test sets and label both with all hashtags in Ht. Rather, we split each Ht into three disjoint sets Htraint, Hvalt, and Htestt according to two time stamps tsplittrain and tsplitval for topic t and the first usage time stamp htime* of each hashtag hHt. In short, all hashtags hHt first used before tsplittrain are used to generate positive labels in the training data, all hashtags hHt first used after tsplittrain and before tsplitval are used to generate positive labels in the validation data, and the remaining hashtags are used to generate positive labels in the test data. Here we first outline the procedure and follow later with a detailed explanation.

To achieve this effect formally, we define the following:

Htraint={h|hHthtime<tsplittrain}

Hvalt={h|hHthtimetsplittrainhtime<tsplitval}

Htestt={h|hHthtimetsplitval}

Once we have split our hashtags into training and validation sets according to tsplittrain and tsplitval, we next proceed to temporally split our training documents D into a training set Dtraint, a validation set Dvalt, and a test set Dtestt for topic t based on the posting time stamp di,time* of each tweet di as follows:

Dtraint={di|diDdi,time<tsplittrain}

Dvalt={di|diDdi,timetsplittraindi,time<tsplitval(hdi:hHtraint)}

Dtestt={di|diDdi,timetvaltrain(hdi:hHtraint)}

Finally, to label the train, validation, and test data sets Dtraint, Dvalt and Dtestt, we use the respective hashtag sets Htraint, Hvalt, Htestt for generating the topic label for a particular tweet t(di) ∈ {0, 1} as follows, where we take a set-based view of the features positively contained in vector di:

t(di)={1ifdiDtrainthdi:hHtraint1ifdiDvalthdi:hHvalt1ifdiDtestthdi:hHtestt0otherwise.

The critical insight here is that we not only divide the train, validation, and test data temporally, but we also divide the hashtag labels temporally and label the validation and test data with an entirely disjoint set of topical labels from the training data. The purpose behind this training, validation and test data split and labeling is to ensure that hyperparameters are tuned so as to prevent overfitting and maximize generalization to unseen topical content (i.e., new hashtags). We remark that by doing this, a classifier that simply memorizes training hashtags will fail to correctly classify the validation data except in cases where a tweet contains both a training and validation hashtag. Moreover, we argue that removing tweets containing training hashtags from the validation data is important since ranking these tweets highly does not provide any indication of classifier generalization beyond the training hashtags. We later experimentally validate this tweet removal proposal against a baseline where (a) we include all train hashtags Htraint in the validation hashtag set Hvalt and (b) we include all tweets di containing these train hashtags in the validation dataset Dvalt.

Per topic, hashtags were split into train and test sets according to their first usage time stamp roughly according to a 3/5 to 2/5 proportion (the test interval spanned between 9–14 months). The train set was further temporally subdivided into train and validation hashtag sets according to a 5/6 to 1/6 proportion. We show a variety of statistics and five sample hashtags per topic in Table 2. Here we can see that different topics had varying prevalence in the data with Soccer being the most tweeted topic and Iran Nuclear Deal being the least tweeted according to our curated hashtags.

Sampling negative examples: Topic classification is often considered to be an imbalanced classification task since usually there are many more negative examples than positive examples. Indeed, the large number of users on Twitter, their diversity, their wide range interests, and the short lifetime of topics discussed on a daily basis typically imply that each topic has only a small set of positive examples. For example, in the “natural disaster” topic that we evaluate in this article, we remark that we have over 800 million negative examples and only 500,000 positive examples. Therefore, given this extreme class imbalance, we have chosen to subsample negative examples, which is commonly used to enable faster training and more effective hyperparameter tuning compared to training with all negative examples. Specifically, we randomly subsample negative examples such that positive examples represent 10% of the dataset for each topic while negative examples represent 90% of the dataset. This rule is valid for the training, validation and test sets of each topic. Detailed statistics of each topic dataset are provided in Table 2.

Training and hyper-parameter tuning: Once Dtraint and Dvalt have been constructed, we proceed to train our scoring function ft on Dtraint and select hyperparameters to optimize Average Precision (AP)2 on Dvalt. Once the optimal ft is found for Dvalt, we return it as our final learned topical scoring function ft for topic t. Because ft(di)R is a scoring function, it can be used to rank.

With train, validation, and testing data defined along with the training methodology, it remains now to extract relevant features, described next.

Topic classification features

The set of features that we consider for each tweet di are: (i) User (author of the tweet), (ii) Mention, (iii) Location, (iv) Term, and (v) Hashtag features. Because we have a total of 538,365,507 unique features in our Twitter corpus (the total count of unique feature values is shown in Table 1), it is critical to pare this down to a size amenable for efficient learning and robust to overfitting. To this end, we thresholded all features according to the frequencies listed in Table 3. The rationale for our frequency thresholding was to have roughly 1 million features with equal numbers of each feature type. We also removed common English stopwords which further reduced the unique term count. Overall, we end up with 1,017,935 candidate features (CF) for learning topical classifiers.

Table 3:
Cutoff threshold and corresponding number of unique values of candidate features CF for learning. Thresholds were chosen to balance the number of each type of feature.
Frequency threshold #Unique values
User 235 206,084
Hashtag 65 201,204
Mention 230 200,051
Location 160 205,884
Term 200 204,712
Total candidate
Features (CF) 1,017,935
DOI: 10.7717/peerj-cs.991/table-3

Supervised learning algorithms

With our labeled training, validation, and test datasets and our candidate feature set CF now defined, we proceed to apply different probabilistic classification and ranking algorithms to generate a scoring function ft for learning topical classifiers as defined previously. In this article, we experiment with the following five state-of-the-art supervised classification and ranking methods:

  1. Logistic Regression (LR) (Fan et al., 2008): LR uses a logistic function to predict the probability that a tweet is topical. We used L2 regularization with the hyperparameter C (the inverse of regularization strength) selected from a search over the values C ∈ {10−12, 10−11, …, 1011, 1012}.

  2. Naïve Bayes (NB) (McCallum & Nigam, 1998): NB makes a naïve assumption that all are features are independent conditioned on the class label. Despite the general incorrectness of this independence assumption, McCallum & Nigam (1998) remark that it is known to make an effective topic classifier. Like LR, NB predicts the probability that a tweet is topical. For parameter estimation, we used Bayesian smoothing using Dirichlet priors with hyperparameter α selected from a search over the values α ∈ {10−20, 10−15, 10−8, 10−3, 10−1, 1}.

  3. RankSVM (Lee & Lin, 2014): RankSVM is a variant of the support vector machine algorithm used to learn from pairwise comparison data (in our case pairs consist of a positive labeled datum that should be ranked above a negatively labeled datum) that naturally produces a ranking. We used a linear kernel with the regularization hyperparameter C (the trade-off between training error and margin) selected in the range C ∈ {10−12, 10−11, …, 1011, 1012}.

  4. Random Forest (RF) (Breiman, 2001): RF is an ensemble learning method for classification that operates by constructing a multitude of decision trees at training time and predicting the class that is the mode of the class prediction of the individual trees (the number of trees that predict the most common class being the score). RF is known to be a classifier that generalizes well due to its robustness to overfitting. For RF, we tuned the hyperparameter for the number of trees in the forest selected from a search over the respective values {10, 20, 50, 100, 200}.

  5. k-Nearest Neighbors (k-NN) (Aha, Kibler & Albert, 1991): k-NN is a non-parametric method used for classification. An instance is classified by a plurality vote of its k neighbors, with the object being assigned to the class most common among its k nearest neighbors (the number of k neighbors for the most common class being the score). The value of k is the primary hyperparameter for k-NN and was selected from a search over the respective values {1, 2, 3, …, 10}.

We remark that almost all algorithms performed better with feature selection and hence we used feature selection for all algorithms, where the number of top features M was selected in a topic-specific manner based on their Mutual Information with the topic being classified. M was tuned over values in {102, 103, 104, 105}. As noted previously, hyperparameter tuning is done via exhaustive grid search using the Average Precision (AP) ranking metric on validation data. All code to process the raw Twitter data and to train and evaluate these classifiers as described above is provided on github (https://github.com/SocialSensorProject/socialsensor).

In the next section, we present results for an intensive evaluation of these classifiers for our longitudinal study of topic classification on the Twitter data previously described.

Results and Discussion

We now report and discuss the main results of our longitudinal study of topic classification on Twitter.

Classification performance analysis

In the following, we first conduct an analysis of the overall classification performance by comparing the classifiers described above, and then, we describe the outcome of a longitudinal classification performance.

Overall classification performance

While our training data is provided as supervised class labels, we remark that topical classifiers are targeted towards individual users who will naturally be inclined to examine only the highest ranked tweets. Hence we believe ranking metrics represent the best performance measures for the intended use case of this work. While RankSVM naturally produces a ranking, all classifiers are score-based, which also allows them to provide a natural ranking of the test data that we evaluate via the following ranking metrics:

  • AP: Average Precision over the ranked list (Manning, Raghavan & Schütze, 2008); the mean over all topics provides the Mean Average Precision (MAP).

  • P@k: Precision at k for k ∈ {10, 100, 1000}.

While P@10 may be a more standard retrieval metric for tasks such as ad-hoc web search, we remark that the short length of tweets relative to web documents makes it more plausible to look at a much larger number of tweets, hence the reason for also evaluating P@100 and P@1000.

Table 4 evaluates our chosen ranking metrics for each topic. Random Forest is the best performing method on average, except for P@1000 where Logistic Regression performed slightly better in the 3rd significant digit. The generally strong performance of Random Forest is due to its robustness to overfitting Breiman (2001). In general, KNN is only slightly worse than Logistic Regression, while Naïve Bayes and RankSVM typically perform worse. Notably, trained classifiers outperform RankSVM on the ranking task thus justifying the use of trained topic classifiers for ranking.

Table 4:
Performance of topical classifier learning algorithms across metrics and topics with the mean performance over all topics shown in the right column with ± 95% confidence intervals.
The best mean performance per metric is shown in bold.
Tennis Space Soccer Iran nuclear deal Human disaster Celebrity death Social issues Natural disaster Epidemics LGBT Mean
LR AP 0.9590 0.6452 0.5036 0.9807 0.6952 0.9293 0.5698 0.9428 0.4005 0.1559 0.6782 ± 0.1724
NB AP 0.5859 0.8471 0.3059 0.9584 0.4224 0.4658 0.5030 0.3518 0.4050 0.1689 0.5014 ± 0.1494
RankSVM AP 0.702 0.840 0.674 0.586 0.603 0.469 0.370 0.248 0.136 0.082 0.471 ± 0.18
RF AP 0.9344 0.9314 0.5509 0.9757 0.6658 0.9571 0.8213 0.8306 0.5154 0.2633 0.7445 ± 0.14764
KNN AP 0.9550 0.7751 0.4739 0.9752 0.598 0.542 0.5078 0.9599 0.5317 0.1774 0.6496 ± 0.1618
LR P@10 1.0 0.2 0.3 1.0 0.5 0.8 0.2 1.0 0.5 0.6 0.61 ± 0.2012
NB P@10 0.1 0.8 0.0 0.9 0.7 0.1 0.0 0.3 0.1 0.0 0.3 ± 0.2225
RankSVM P@10 1.0 0.8 0.6 0.8 0.4 0.3 0.0 0.1 0.0 0.2 0.42 ± 0.26
RF P@10 1.0 0.5 0.5 1.0 0.9 1.0 1.0 1.0 0.7 0.5 0.81 ± 0.1444
KNN P@10 1.0 0.0 1.0 1.0 0.7 0.9 0.0 0.9 0.3 0.4 0.62 ± 0.2543
LR P@100 0.98 0.65 0.44 0.99 0.74 0.94 0.59 0.98 0.45 0.2 0.696 ± 0.1721
NB P@100 0.56 0.95 0.0 0.98 0.39 0.36 0.16 0.37 0.48 0.1 0.435 ± 0.2033
RankSVM P@100 0.73 0.72 0.31 0.70 0.88 0.44 0.48 0.34 0.02 0.100 0.472 ± 0.20
RF P@100 0.98 0.94 0.43 0.98 0.62 0.97 0.81 0.9 0.61 0.29 0.753 ± 0.1555
KNN P@100 1.0 0.59 0.34 1.0 0.72 0.54 0.39 0.96 0.54 0.24 0.632 ± 0.1731
LR P@1000 0.653 0.703 0.545 0.299 0.666 0.884 0.574 0.919 0.267 0.076 0.5586 ± 0.1682
NB P@1000 0.551 0.667 0.29 0.333 0.338 0.542 0.655 0.287 0.319 0.169 0.4151 ± 0.1073
RankSVM P@1000 0.799 0.922 0.764 0.218 0.525 0.547 0.215 0.173 0.154 0.064 0.438 ± 0.22
RF P@1000 0.728 0.464 0.576 0.331 0.463 0.914 0.789 0.728 0.397 0.159 0.5549 ± 0.145
KNN P@1000 0.571 0.821 0.53 0.329 0.476 0.84 0.49 0.929 0.234 0.083 0.5303 ± 0.1696
DOI: 10.7717/peerj-cs.991/table-4

To provide more insight into the general performance of our learning topical classifier framework, we provide the top five tweets for each topic according to Logistic Regression3 in Table 5. We have annotated tweets with symbols as follows:

Table 5:
Top tweets for each topic from Logistic Regression method results, marked with ✗ as irrelevant, ✓ as relevant and labeled as topical, and ★ as relevant but labeled as non-topical (a mislabeled example).
Tennis Space
✓ PHOTOS; @andy_murray in @usopen QF match v Novak Djokovic … @usta @BritishTennis #USOpen2014… ✗ RT @wandakki: Chuck’s Story - My 600-lb Life — http://t.co/aP3L1OqIch — Reality TV #tv #episode #Reality #TV…
✓ PHOTOS; British #1 @andy_murray in @usopen Quarter-Finals match v Novak Djokovic … @usta @BritishTennis #USOpen2014… ✗ RT @arist_brain: Path. #Switzerland (by Roman Burri) #travel #landscape #nature #path #sky #alps #clouds…
✓ RT @fi_sonic: PHOTOS; @andy_murray in 75 75 64 win over Jo-wilfried Tsonga to reach @usopen QFs. @BritishTennis… ✗ TeamFest Winner Circle by Dee n Ralph on Etsy--Pinned with http://t.co/Cr1PC31naR #beach #ocean #sea #love…
✓ PHOTOS; #21 seed @sloanetweets in her @usopen 2nd round match v Johanna Larsson … @USTA @WTA #USOpen2014… ✓ RT @NASA: Fire @YosemiteNPS as seen by NASA’s Aqua satellite on Sunday. #EarthRightNow…
✓ “ @fi_sonic: PHOTOS; @DjokerNole celebrating his @usopen QF match win 76 67 62 64 v Andy Murray … @usta #USOpen2014… ✓ RT @NASA: Arkansas April 27 tornado track seen by NASA’s EO-1 satellite. http://t.co/d36sKPGzAx #EarthRightNow…
Soccer Iran Nuclear Deal
✓ RT @FOXSoccer: Cameron in for Beckerman #USA lineup: Howard, Gonzalez, Bradley, Besler, Beasley, Dempsey… ✓ RT @JavadDabiran: #Iran-Executions, #Women rights abuse, #IranHRviolations soar under Hassan Rouhani #No2Rouhani…
✓ RT @FOXSoccer: Cameron in for Beckerman #USA lineup: Howard, Gonzalez, Bradley, Besler, Beasley, Dempsey ✓ RT @HellenaRezai: #Iran-Executions, #Women rights abuse, #IranHRviolations soar under Hassan Rouhani #No2Rouhani…
★ RT @Gerrard8FanPage: Luis Suarez has scored seven goals in six Barclays Premier League appearances against Sunderland. ✓ RT @peymaneh123: #Iran-Executions, #Women rights abuse, #IranHRviolations soar under Hassan Rouhani #No2Rouhani…
★ RT @BBCMOTD: Federico Fazio is the first player sent off on his PL debut since Samba Diakite for #QPR in Feb 2012 #THFC… ✓ RT @IACNT: #Iran nuclear threat bigger than claimed: http://t.co/13Qk7cyWyA @SenTedCruz @JohnCornyn #nuclear…
★ @JamesYouCun* well I’d say Migs, moreno sakho toure (if fit) manquilio Lucas can gerrard sterling Coutinho markovic and borini ✓ RT @YelloJackets: #Iran-Executions, Women rights abuse and #IranHRviolations soar under Hassan Rouhani
Human Disaster Celebrity Death
✓ @IlenePrusher if one thinks of Gazan kids as potential Hamas fighters Gazan women as potential Hamas fighters’ mothers, yes! ✓ #RIPRise Heaven gained another angel yet another angel, you will be happy with EunB, all our prayers are for you…
★ RT @jala_leb: This is GAZA not Hiroshima @BarackObama @David_Cameron @un @hrw http://t.co/ddZWORPqrQ ✓ RT @WeGotLoves: EunB, Manager, Driver Rise passed away. Very heartbreaking news. Deep condolences to their family.
✓ RT @jallubk: THIS AGAIN: BOYCOTT ISRAEL OR WE WILL BOYCOTT YOU, @robbiewilliams ! #IsraelKillsKids… ✓ RT @sehuntella: eunb, manager, driver and rise passed away. what a heartbreaking news. deep condolences to their family
✗ RT @notdramadriven: Nailed it @KenWahl1 @DrMartyFox @jjauthor @shootingfurfun @CarmineZozzora… ✓ RT @missA_TH: Our deep condolences to family, friends and fans of EunB Rise. May they rest in peace. Heaven has
✓ RT @TelecomixCanada: @Op_Israel #Article51 of the Geneva Convention: http://t.co/VaDklflx5C Tick Tock ✓ Rest in peace Rise! Heaven now gained two angels. #RipRise #PrayForLadiesCode My condolences :(
Social Issues Natural Disaster
✓ RT @RightCandidates: THANK YOU DEMOCRAT RACE BAITERS #tcot #america #women #millennials #tlot… ✓ RT @ianuragthakur: I appeal to friends supporters @BJYM to help in the relief efforts fr #KashmirFloods…
✗ RT @2AFight: The Bill of Rights IS my Patriot Act #2A #NRA #MolonLabe #RKBA #ORPUW #PJNET #tgdn… ✓ RT @RSS_Org: RSS Press Release: An Appeal to the Society to donate for Relief Fund to help #KashmirFloods Victims…
✗ The Supreme Court Judicial Tyranny http://t.co/HKo4hnQnF5 #1A #MakeDCListen #NObamaCARE#KeystoneXL… ✓ RT @punkboyinsf: #BREAKING California Gov. Jerry Brown has declared a state of emergency following…
✓ RT @RightCandidates: THANK YOU DEMOCRAT RACE BAITERS FOR THIS #tcot #america #women #FergusonDecision… ✓ RT @nbcbayarea: #BREAKING California Gov. Jerry Brown has declared a state of emergency following…
✗ Race-Baiting for Profit RT http://t.co/KOYfDDNQCu #TCOT #CCOT #MakeDCListen #TeaParty #Conservatives ✓ RT @coolfunnytshirt: Congress ke bure din! RT @timesnow: Congress leader Saifuddin Soz heckled by flood victims
Epidemics LGBT
✓ RT @justgrateful: Surgeon General Nominee is Blocked by NRA #occupy #uppers #tcot #ccot #topprog #EbolaCzar… ✗ RT @CSGV: Take a bite out of the crime. Oppose traitors preparing for war w/ our gov’t. #NRA #NRAAM Cliven Bundy…
✓ RT @nhdogmom: Why don’t we have Surgeon General/Medical #EbolaCzar … GOP RWNJ’s is why!!… ✗ IRS employee suspended for pro-Obama… - Washington Times: http://t.co/KoCtwaJ0C6via @washtimes Another meaningless…
✗ New York seen like never before! #cool #photo #black white #atmospheric #moody ✓ Pa. gay-marriage ban overturned http://t.co/Gl4kAhQwyQvia @phillydotcom #lovewins #lgbt
✗ RT @ryangrannand:.@CouncilW9 asking developer for a sign plan. #waltham ✓ RT @OR4Marriage: RT this AMAZING quote from yesterday’s ruling striking down #Oregon’s marriage ban! #OR4M #lgbt…
✗ GOOD OFFER!! http://t.co/1qm1K0UIaw Vitamins Supplements, Clinically Proven-Doctor Formulated ✓ @briansbrown YOU ANTI-GAY BIGOTS ARE BOX-OFFICE-POISON EVEN FOR MOST REPUBLICANS. #LGBT…
DOI: 10.7717/peerj-cs.991/table-5
  • ✓: the tweet was labeled topical by our test hashtag set.

  • ★: the tweet was determined to be topical through manual evaluation even though it did not contain a hashtag in our curated hashtag set (this corresponds to a mislabeled example due to the non-exhaustive strategy used to label the data).

  • ✗: the tweet was not topical.

In general, we remark that our topical classifier may perform slightly better than the quantitative results in Table 4 would indicate: a few of the highly ranked tweets are mislabeled as non-topical in the test set although a manual analysis reveals that they are in fact topical. Furthermore, even though we use hashtags to label our training, validation, and testing data, our topical classifier has highly (and correctly) ranked topical tweets that do not contain hashtags, indicating strong generalization properties from a relatively small set of curated topical hashtags.

Though the reason why some non-topical tweets ranked highly is unclear, we see that many failure cases appear to mention relevant features to the topic although they are in fact advertising or politicized spam content. This indicates a limitation of the hashtag-based class labeling method, which cannot easily distinguish spam from legitimate content. Nonetheless, we believe that a separate spam filter common to all classifiers could mitigate these issues since the patterns of spam email such as an unusually large number of hashtags or mentions are not topic-specific and can be easily detected.

Longitudinal classification performance

Now that we’ve examined the overall classification performance of different topical classifiers per topic and per metric, we now turn to address the long-term temporal aspect of topic classification with two questions: (1) Does classification performance degrade as time increases since training, and if so, by how much? (2) Does omission of training hashtags from the validation set encourage better long-term generalization since, as hypothesized in the methodology, it discourages memorizing training hashtags?

To assess these questions, Figs. 2A2D plots the performance of the Logistic Regression4 topic classifier (mean over all 10 topics) from 50 to 350 days after training, evaluated according to (a) mean AP (MAP), (b) P@10, (c) P@100, and (d) P@1000. The purple line shows the proposed methodology, where tweets with training hashtags are suppressed from the validation set, while the green line does not suppress training hashtags (see the Methodology section for more details on both methods). To better distinguish the overall performance of suppressing training hashtags in the validation set, we average results over all time points in Fig. 2E.

Longitudinal analysis of classifier generalization.
Figure 2: Longitudinal analysis of classifier generalization.
(A–D) The performance of the topic classifier (mean over all 10 topics with 95% confidence intervals) from 50 to 350 days after training, evaluated according to (A) mean AP (MAP), (B) P@10, (C) P@100, and (D) P@1000. Best fit linear regressions are shown as dashed lines. (E) Results averaged over time with 95% confidence intervals.

Overall, we make a few key observations:

  • Regarding question (1), it is clear that the classification performance drops over time–a roughly 35% drop in MAP from the 50th to the 350th day after training. Clearly, there will be topical drift over time for most topics (e.g., Natural Disasters, Social Issues, Epidemics) as different events occur and shift the focus of topical conversation. While there are more sophisticated training methods for mitigating some of this temporal drift (e.g., Wang et al., 2019), overall, it would seem that the most practical and effective method for long-term generalization would involve a periodic update of training hashtags and data labels.

  • Regarding question (2), Fig. 2E clearly shows an overall performance improvement from discarding training hashtags (and their tweets) from the validation set. In fact, for MAP alone, we see roughly an 11% improvement. Hence, these experiments suggest there may be a long-term generalization advantage to excluding training hashtags from the validation hashtags and data, which we conjecture discourages hyperparameters that lead to hashtag memorization from the training set.

With our comparative and longitudinal analysis of topic classifier performance now complete, we will next investigate which features are most informative for topic classifiers.

Feature analysis

In this section, we analyze the informativeness of feature sets defined in the Data Description section and the effect of their attributes on learning targeted topical classifiers. To this end, our goal in this section is to answer the following questions:

  • What are the best features for learning classifiers and do they differ by topic?

  • For each feature type, do any attributes correlate with importance?

To answer these questions, we use Mutual Information (MI) (Manning, Raghavan & Schütze, 2008) as our primary metric for feature evaluation. MI is a general method for measuring the amount of information one random variable contains about another random variable and is used to select predictive features in machine learning. To calculate the amount of information that each feature j in the Candidate Features (CF) defined previously provides w.r.t. each topic label t ∈ {Natural Disaster, Epidemics, …}, MI is formally defined as

I(j,t)=t{0,1}j{0,1}p(j,t)log(p(j,t)p(j)p(t))with marginal probabilities of topic p(t) and feature p(j) occurrence and joint probability p(t, j) computed empirically over the sample space of all tweets, where higher values for this metric indicate more informative features j for the topic t.

In order to assess the overall best feature types for learning topical classifiers, we provide the mean MI values for each feature type across different topics in Fig. 3. The last column in Fig. 3 shows the average of the mean MI for each feature type and the last row shows the average of the mean MI for each topic. From analysis of Fig. 3, we make the following observations:

Matrix of mean Mutual Information values for different feature types vs. topics.

Figure 3: Matrix of mean Mutual Information values for different feature types vs. topics.

The last column and last row represent the average of mean values across all topics and all features respectively. All values should be multiplied by 10−8.
  • Looking at the average MI values, the order of informativeness of feature types is the following: Hashtag, Term, Mention, User, Location. The overall informativeness of Hashtags is not surprising given that hashtags are used on Twitter to tag topics of interest. While the Term feature is not strictly topical, it contains a rich vocabulary for describing topics that Mention, User, and Location lack.

  • The Location feature provides high MI regarding the topics of Human Disaster, LBGT, and Soccer indicating that a lot of content in these topics is geographically localized.

  • Revisiting Table 4, we note the following ranking of topics from highest to lowest AP for Logistic Regression5 : Iran, Tennis, Natural Disaster, Celebrity Death, Human Disaster, Space, Social Issue, Soccer, Epidemics, LGBT. It turns out that this ranking is anti-correlated with the ranking of topics according to average MI of features in Fig. 3. To establish this relationship more clearly, in Fig. 4 we show a scatterplot of topics according to MI rank vs. AP rank. Clearly, we observe that there is a negative correlation between the topic ranking based on AP and MI; in fact, the Kendall τ rank correlation coefficient is −0.68 indicating a fairly strong inverse ranking relationship. To explain this, we conjecture that lower average MI indicates that there are fewer good features for a topic; however, this means that classifiers for these topics can often achieve high ranking precision because there are fewer good features and the tweets with those features can be easily identified and ranked highly, leading to high AP. The inverse argument should also hold.

Scatter plot showing ranking of topics w.r.t. Mutual Information vs. Average Precision.

Figure 4: Scatter plot showing ranking of topics w.r.t. Mutual Information vs. Average Precision.

There is clearly a negative correlation, with a Kendall τ coefficient of −0.68.

To further analyze the relationship between the informativeness of feature types and topics, we refer to the box plots of Fig. 5. Here we see the quartiles and outliers of the distribution rather than just the average of the MI values in order to ensure the mean MI values were not misleading our interpretations. Overall, the story of feature informativeness becomes much more complex, with key observations as follows:

Box plots of Mutual Information values (y-axis) per feature type across topics (x-axis labels).

Figure 5: Box plots of Mutual Information values (y-axis) per feature type across topics (x-axis labels).

  • The topic has little impact on which feature is most important, indicating stability of feature type informativeness over topics.

  • While Hashtag had a higher mean MI score than Term in the previous analysis, we see that Term has the highest median MI score across all topics, indicating that the high mean MI of Hashtag is mainly due to its outliers. In short, the few good Hashtag outliers are the overall best individual features, while Term has a greater variety of strong (but not absolute best) features.

  • Across all topics, User is often least informative. However, the distribution of Location and Mention typically performs competitively with Hashtag, although their outliers do not approach the best Hashtag features, explaining why Hashtag has an overall higher average in Fig. 3.

Now we proceed to a more nuanced analysis of feature types for each topic according to the proportions of their presence among the top p% percentiles of MI values for p% ∈ {0.001%, 0.01%, 0.1%, 1%, 10%} as shown in Fig. 6. Here we make a few key observations:

Top p% features ranked by Mutual Information.

Figure 6: Top p% features ranked by Mutual Information.

  • Overall, Hashtags dominate the top 0.001 percentile of features indicating that they account for the most informative features overall.

  • However, from percentiles 0.01 to 10, we largely see an increasing proportion of Term features among each percentile. This indicates that while the most informative features are Hashtags, there are relatively few of them compared to the number of high MI terms.

  • Not to the same extent as Terms, we note that Mentions also start to become notably more present as the percentile range increases, while Locations and Users appear least informative overall among the 10th percentile and smaller.

As anecdotal evidence to inspect which features are most informative, we refer to Table 6, which displays the top five feature instances according to MI for each feature type and topic. For example the term typhoon is the highest MI term feature with the topic Natural Disaster, the official UNICEF6 Twitter account (@unicef) is the highest MI feature mention with the Human Disaster topic, and #worldcup is (unsurprisingly) the highest MI hashtag feature for the topic Soccer. The top locations are also highly relevant to most topics indicating the overall importance of these tweet features for identifying topical tweets; for example, three variations of St. Louis, Missouri appear as top MI locations for topic Social Issues7 . One general observation is that Hashtag and Term features are appear to be the most generic (and hence most generalizable) features, providing strong intuition as to why these features figure so prominently in terms of their informativeness8 .

Table 6:
The top five features for each feature type and topic based on Mutual Information.
Topics/Top10 Natural disaster Epidemics Iran deal Social issues LBGT Human disaster Celebrity death Space Tennis Soccer
User from_japan changedecopine mazandara debtadvisoruk stevendickinson witfp boiknox daily_astrodata tracktennisnews makeupbella
User everyearthquake stylishoz freeiran9292 nsingerdebtpaid mgdauber ydumozyf jacanews freesolarleads novakdjokovic_i sport_agent
User quakestoday drdaveanddee hhadi119 negativeequityf lileensvf1 syriatweeten ewnreporter sciencewatchout i_roger_federer yasmingoode
User equakea soliant_schools balouchn2 iris_messenger kevinwhipp rk70534 rowwsupporter houston_jobs andymurrayfans1 sportsroadhouse
User davewinfields msgubot jeffandsimon dolphin_ls petermabraham gosyrianews flykiidchris lenautilus rafaelnadal_fan losangelessrh
Hashtag #earthquake #health #iran #ferguson #tcot #syria #rip #science #wimbledon #worldcup
Hashtag #haiyan #uniteblue #irantalks #mikebrown #pjnet #gaza #ripcorymonteith #sun #tennis #lfc
Hashtag #storm #ebola #iranian #ericgarner #p2 #israel #riprobinwilliams #houston #usopen #football
Hashtag #PrayForThePhilippines #healthcare #rouhani #blacklivesmatter #uniteblue #gazaunderattack #rippaulwalker #starwars #nadal #worldcup2014
Hashtag #tornado #fitness #irantalksvienna #icantbreathe #teaparty #isis #robinwilliams #scifi #wimbledon2014 #sports
Location With everyone USA France St Louis MO USA Syria South Africa Houston TX Worldwide Liverpool
Location Earth Francophone Tehran Iran Washington DC Bordentown New Jersey Palestine Pandaquotescom Germany London Manchester
Location Philippines United States Inside of Iran St Louis Global Markets Syrian Arab Republic Johannesburg South Africa Houston The Midlands London
Location Don’t follow me am i a bot Gainesville FL USA Iran Virginia US The blue regime of Maryland Israel Johannesburg Rimouski London UK Anfield
Location Global planet earth Boulder Colorado Washington DC Saint Louis MO Lancaster county PA Washington DC Cape Town In a galaxy far far ebay Wimbledon Bangil East Java Indonesia
Mention @oxfamgb @foxtramedia @ap @natedrug @jjauthor @ifalasteen @nelsonmandela @nasa @wimbledon @lfc
Mention @gabriele_corno @obi_obadike @afp @deray @2anow @drbasselabuward @realpaulwalker @philae2014 @usopen @fifaworldcup
Mention @weatherchannel @who @iran_policy @antoniofrench @gop @revolutionsyria @ddlovato @maximaxoo @atpworldtour @ussoccer
Mention @twcbreaking @kayla_itsines @4freedominiran @bipartisanism @pjnet_blog @unicef @robinwilliams @esa_rosetta @andy_murray @mcfc
Mention @redcross @canproveit @orgiac @theanonmessage @espuelasvox @free_media_hub @historicalpics @astro_reid @wta @realmadriden
Term typhoon health nuclear police obama israeli robin space tennis liverpool
Term philippines ebola regime protesters gun israel williams solar murray cup
Term magnitude outbreak iran officer america gaza walker moon djokovic supporting
Term storm virus iranian cops obamacare palestinian cory houston federer match
Term usgs acrx mullahs protest gop killed paul star nadal goal
DOI: 10.7717/peerj-cs.991/table-6

In order to answer the second question on whether any attributes correlate with importance for each feature, we provide two types of analysis using the topic Celebrity Death–the other topics showed similar patterns, thus we have chosen to omit them. The first analysis shown in Fig. 7 analyzes the distributions of Mutual Information values for features when binned by the magnitude of various attributes of those features, outlined as follows:

Boxplots for the distribution of Mutual Information values (y-axis) of different features as a function of their attribute values (binned on x-axis).

Figure 7: Boxplots for the distribution of Mutual Information values (y-axis) of different features as a function of their attribute values (binned on x-axis).

Plots (A–E) respectively show attributes {favorite count, follower count, friend count, hashtag count, tweet count} for From feature. Plots (F–J) respectively show attributes tweetCount and userCount for Hashtag, userCount for Location feature, tweetCount for Mention and Term features.
  • Uservs.

  • Favorite count: # of tweets user has favorited.

  • Followers count: # of users who follow user.

  • Friends count: # of users followed by user.

  • Hashtag count: # of hashtags used by user.

  • Tweet count: # of tweets from user.

  • Hashtagvs.

  • Tweet count: # of tweets using hashtag.

  • User count: # of users using hashtag.

  • Locationvs. User count: # of users using location.

  • Mentionvs. Tweet count: # of tweets using mention.

  • Termvs. Tweet count: # of tweets using term.

As we can see in the boxplots of Fig. 7, the general pattern is that the greater the number of tweets, users, or hashtag count a feature has, the more informative the feature is in general. This pattern also exists to some extent on the attributes of the From feature, although the pattern is less visible in general and not clear (or very weak) for the follower or friend count. In general, the informativeness of a user appears to have little correlation with their follower or friend count.

Figure 8 provides a further analysis by showing density plots of the tweet count attribute of the User, Hashtag, Mention and Term features, and the user count attribute of the Hashtag feature. Here we can clearly observe the positive linear correlation that exists between the attribute magnitude and the Mutual Information value for all of the evaluated attributes. In short, the more tweets using User, Hashtag, Mention and Term features and the more users using a Hashtag feature, the more informative that feature typically is for the topic.

Density plots for the frequency values of feature attributes (x-axis) vs. Mutual Information (y-axis).

Figure 8: Density plots for the frequency values of feature attributes (x-axis) vs. Mutual Information (y-axis).

Plots (A–E) respectively show the following attributes: number of tweets for the User feature, number of tweets for the Hashtag feature, number of users using the Hashtag feature, number of tweets for the Mention feature, and number of tweets for the Term feature.

Conclusions

This work provides a long-term study of topic classifiers on Twitter that further justifies classification-based topical filtering approaches while providing detailed insight into the feature properties most critical for topic classifier performance. Our results suggest that these learned topical classifiers generalize well to unseen future topical content over a long time horizon (i.e., 1 year) and provide a novel paradigm for the extraction of high-value content from social media. Furthermore, an extensive analysis of features and feature attributes across different topics has revealed key insights including the following two: (i) largely independent of topic, hashtags are the most informative features followed by generic terms, and (ii) the number of unique hashtags and tweets by a user correlates more with their informativeness than their follower or friend count.

Among many interesting directions, future work might evaluate a range of topical classifier extensions: (1) optimizing rankings not only for topicality but also to minimize the lag-time of novel content identification, (2) optimizing queries for boolean retrieval oriented APIs such as Twitter, (3) identification of long-term temporally stable predictive features, (4) utilizing more social network structure as graph-based features, and (5) investigating classifier performance based on topic properties such as periodicity over time or specificity to a very narrow audience. Altogether, we believe these insights will facilitate the continued development of effective topical classifiers for Twitter that learn to identify broad themes of topical information with minimal user interaction and enhance the overall social media user experience.

This is an extended and revised version of a preliminary conference report that was presented in Iman et al. (2017).
See Manning, Raghavan & Schütze (2008) for a discussion and definition of this commonly used ranking metric.
Logistic Regression allows us to better understand failure cases for topical classifiers, i.e., Random Forest is likely to have gotten all of the top-5 right.
We could not run these longitudinal experiments with Random Forest due to the significant computational expense of the analysis in this section and the hyperparameter tuning that is required, thus we opted to perform this analysis with the much faster and still strongly competitive Logistic Regression classifier.
The ranking for Random Forest only differs slightly.
The United Nations Children’s Fund (UNICEF) is an organization that aims to provide emergency food and healthcare to children and mothers in developing countries everywhere.
We remark that the original Black Lives Matter protests originated in St. Louis, Missouri in the aftermath of the police shooting of Michael Brown on August 9, 2014.
It should also be remarked that Mutual Information (MI) is very sensitive to frequency so a high MI feature must be both informative and frequent to rank highly. This explains why the high MI features are so generic, i.e., they are frequent and hence cover many more tweets than low MI features.
1 Citation   Views   Downloads