Sreekanth Madisetty, PhD

Twitter provides access to latest information. Whenever a major event happens, people try to search for event related information in social media platforms like Twitter. So, it is essential to develop methods to get good quality of event related tweets. People share different opinions, feelings, feedback, etc. about events happening around the world in Twitter in the form of tweets. These tweets are often short and contain noise. So, it is very difficult to get the most relevant data for a… Show more

Twitter provides access to latest information. Whenever a major event happens, people try to search for event related information in social media platforms like Twitter. So, it is essential to develop methods to get good quality of event related tweets. People share different opinions, feelings, feedback, etc. about events happening around the world in Twitter in the form of tweets. These tweets are often short and contain noise. So, it is very difficult to get the most relevant data for a given event from Twitter. We propose a two-phase approach to retrieve the tweets related to planned events. In the first phase, initial retrieval is done by using BM25 algorithm. In the second phase, reranking is done by combining three scoring mechanisms namely BM25 score, top hashtags score related to an event, and top TF-IDF terms score related to an event. A learning to rank algorithm SVM_Rank is applied to give weights to these three methods and combine them to get the final score of the tweet. We performed experiments on two benchmark datasets CLEF and TREC. Experimental results show that our method outperforms baseline and literature methods for both the datasets according to multiple evaluation metrics. Show less

Early prediction of popularity is crucial for recommendation of planned events such as concerts, conferences, sports events, performing arts, etc. Estimation of the volume of social media discussions related to the event can be useful for this purpose. Most of the existing methods for social media popularity prediction focus on estimating tweet popularity i.e. predicting the number of retweets for a given tweet. There is less focus on predicting event popularity using social media. We focus on… Show more

Early prediction of popularity is crucial for recommendation of planned events such as concerts, conferences, sports events, performing arts, etc. Estimation of the volume of social media discussions related to the event can be useful for this purpose. Most of the existing methods for social media popularity prediction focus on estimating tweet popularity i.e. predicting the number of retweets for a given tweet. There is less focus on predicting event popularity using social media. We focus on predicting the popularity of an event much before its start date. This type of early prediction can be helpful in event recommendation systems, assisting event organizers for better planning, dynamic ticket pricing, etc. We propose a deep learning based model to predict the social media popularity of an event. We also incorporate an extra feature indicating how many days left to the event start date to improve the performance. Experimental results show that our proposed deep learning based approach outperforms the baseline methods. Show less

Scientific documents generally contain multiple mathematical expressions in them. Detect-
ing inline mathematical expressions is one of the most important and challenging tasks in
scientific text mining. Recent works that detect inline mathematical expressions in scientific documents have looked at the problem from an image processing perspective. There is
little work that has targeted the problem from NLP perspective. Towards this, we define
a few features and applied Conditional… Show more

Scientific documents generally contain multiple mathematical expressions in them. Detect-
ing inline mathematical expressions is one of the most important and challenging tasks in
scientific text mining. Recent works that detect inline mathematical expressions in scientific documents have looked at the problem from an image processing perspective. There is
little work that has targeted the problem from NLP perspective. Towards this, we define
a few features and applied Conditional Random Fields (CRF) to detect inline mathemat-
ical expressions in scientific documents. Apart from this feature based approach, we also
propose a hybrid algorithm that combines Bidirectional Long Short Term Memory net-
works (Bi-LSTM) and feature-based approach for this task. Experimental results suggest
that this proposed hybrid method outperforms several baselines in the literature and also
individual methods in the hybrid approach. Show less

Every day, lots of events keep happening around the globe. Examples of events are festivals, concerts, shows, conferences, sports events, movie launches, etc. There are several event recommendation systems that suggest different events to the users. The popularity of an event plays an important role in event recommendation. If the system knows which event is popular in a region, then that can be recommended to more people in that region. However, since events take place at predefined location… Show more

Every day, lots of events keep happening around the globe. Examples of events are festivals, concerts, shows, conferences, sports events, movie launches, etc. There are several event recommendation systems that suggest different events to the users. The popularity of an event plays an important role in event recommendation. If the system knows which event is popular in a region, then that can be recommended to more people in that region. However, since events take place at predefined location and time it is important to identify the popularity much before the occurrence of the event. Early prediction of future popularity is very crucial for the success of any event recommendation system. We plan to use event related discussions in social media as a signal for estimating the popularity of the events. Towards this task, given meta information (time, date, venue, title etc.) about an event, we first try to identify social media posts related to the event. These posts are then passed through several preprocessing stages in a pipeline to detect spam, identify the presence of emotions and their intensities, etc. The preprocessed posts can then be used along with several other contextual parameters to predict the future popularity of the events. Next, events can be recommended to the users by using a recommendation algorithm. Show less

Lots of planned events (e.g. concerts, sports matches, festivals, etc.) keep happening across the world every day. In various applications like event recommendation, event reporting, etc. it might be useful to find user discussions related to such events from social media. Identification of event related hashtags can be useful for this purpose. In this paper, we focus on identifying the top hashtags related to a given event. We define a set of features for (event, hashtag) pairs, and discuss… Show more

Lots of planned events (e.g. concerts, sports matches, festivals, etc.) keep happening across the world every day. In various applications like event recommendation, event reporting, etc. it might be useful to find user discussions related to such events from social media. Identification of event related hashtags can be useful for this purpose. In this paper, we focus on identifying the top hashtags related to a given event. We define a set of features for (event, hashtag) pairs, and discuss ways to obtain these feature scores. A linear aggregation of these scores is used to finally output a ranked list of top hashtags for the event. The aggregation weights of the features are obtained using a learning to rank algorithm. We establish the superiority of our method by performing detailed experiments on a large dataset containing multiple categories of events and related tweets. Show less

As the social networking sites get more popular, spammers target these sites to spread spam posts. Twitter is one of the most popular online social networking sites where users communicate and interact on various topics. Most of the current spam filtering methods in Twitter focus on detecting the spammers and blocking them. However, spammers can create a new account and start posting new spam tweets again. So there is a need for robust spam detection techniques to detect the spam at tweet… Show more

As the social networking sites get more popular, spammers target these sites to spread spam posts. Twitter is one of the most popular online social networking sites where users communicate and interact on various topics. Most of the current spam filtering methods in Twitter focus on detecting the spammers and blocking them. However, spammers can create a new account and start posting new spam tweets again. So there is a need for robust spam detection techniques to detect the spam at tweet level. These types of techniques can prevent the spam in real time. To detect the spam at tweet level, often features are defined, and appropriate machine learning algorithms are applied in the literature. Recently, deep learning methods are showing fruitful results on several natural language processing tasks. We want to use the potential benefits of these two types of methods for our problem. Toward this, we propose an ensemble approach for spam detection at tweet level. We develop various deep learning models based on convolutional neural networks (CNNs). Five CNNs and one feature-based model are used in the ensemble. Each CNN uses different word embeddings (Glove, Word2vec) to train the model. The feature-based model uses content-based, user-based, and n-gram features. Our approach combines both deep learning and traditional feature-based models using a multilayer neural network which acts as a meta-classifier. We evaluate our method on two data sets, one data set is balanced, and another one is imbalanced. The experimental results show that our proposed method outperforms the existing methods. Show less

With the rise of user-generated content in social media coupled with almost non-existent moderation in many such systems, aggressive contents have been observed to rise in such forums. In this paper, we work on the problem of aggression detection in social media. Aggression can sometimes be expressed directly or overtly or it can be hidden or covert in the text. On the other hand, most of the content in social media is non-aggressive in nature. We propose an ensemble based system to classify an… Show more

With the rise of user-generated content in social media coupled with almost non-existent moderation in many such systems, aggressive contents have been observed to rise in such forums. In this paper, we work on the problem of aggression detection in social media. Aggression can sometimes be expressed directly or overtly or it can be hidden or covert in the text. On the other hand, most of the content in social media is non-aggressive in nature. We propose an ensemble based system to classify an input post to into one of three classes, namely, Overtly Aggressive, Covertly Aggressive, and Non-aggressive. Our approach uses three deep learning methods, namely, Convolutional Neural Networks (CNN) with five layers (input, convolution, pooling, hidden, and output), Long Short Term Memory networks (LSTM), and Bi-directional Long Short Term Memory networks (Bi-LSTM). A majority voting based ensemble method is used to combine these classifiers (CNN, LSTM, and Bi-LSTM). We trained our method on Facebook comments dataset and tested on Facebook comments (in-domain) and other social media posts (cross-domain). Our system achieves the F1-score (weighted) of 0.604 for Facebook posts and 0.508 for social media posts. Show less

With the increased popularity of online social networks, spammers find these platforms easily accessible to trap users in malicious activities by posting spam messages. In this work, we have taken Twitter platform and performed spam tweets detection. To stop spammers, Google SafeBrowsing and Twitter's BotMaker tools detect and block spam tweets. These tools can block malicious links, however they cannot protect the user in real-time as early as possible. Thus, industries and researchers have… Show more

With the increased popularity of online social networks, spammers find these platforms easily accessible to trap users in malicious activities by posting spam messages. In this work, we have taken Twitter platform and performed spam tweets detection. To stop spammers, Google SafeBrowsing and Twitter's BotMaker tools detect and block spam tweets. These tools can block malicious links, however they cannot protect the user in real-time as early as possible. Thus, industries and researchers have applied different approaches to make spam free social network platform. Some of them are only based on user-based features while others are based on tweet based features only. However, there is no comprehensive solution that can consolidate tweet's text information along with the user based features. To solve this issue, we propose a framework which takes the user and tweet based features along with the tweet text feature to classify the tweets. The benefit of using tweet text feature is that we can identify the spam tweets even if the spammer creates a new account which was not possible only with the user and tweet based features. We have evaluated our solution with four different machine learning algorithms namely - Support Vector Machine, Neural Network, Random Forest and Gradient Boosting. With Neural Network, we are able to achieve an accuracy of 91.65% and surpassed the existing solution by approximately 18%. Show less

Recently, user generated contents have increased tremendously in social media. Twitter is a popular micro-blogging platform in which users share their feelings, opinions, feedback, etc. It has been observed that microblogs are often associated with emotions. Several studies have focused on assigning a given tweet to one of the available emotion categories (e.g., anger, fear, joy, sadness). It is often useful in applications to find the intensity of emotion in the tweets. The focus on… Show more

Recently, user generated contents have increased tremendously in social media. Twitter is a popular micro-blogging platform in which users share their feelings, opinions, feedback, etc. It has been observed that microblogs are often associated with emotions. Several studies have focused on assigning a given tweet to one of the available emotion categories (e.g., anger, fear, joy, sadness). It is often useful in applications to find the intensity of emotion in the tweets. The focus on identifying emotion intensity is less in the literature. In this paper, we focus on determining the level of emotion intensity in the tweets. We use an ensemble of three methods: Convolution Neural Networks (CNN) with word embedding features, XGBoost with word n-gram and char n-gram features, and Support Vector Regression (SVR) with lexicon and word embedding features. The final prediction of the given tweet is obtained by the average of predictions of individual methods in the ensemble. The performance of ensemble is better than the methods in the ensemble due to diverse features. Our experimental results outperform baseline methods. Show less

Users in social media often participate in discussions regarding different events happening in the physical world (e.g., concerts, conferences, festivals) by posting messages, replying to or forwarding messages related to such events. In various applications like event recommendation, event reporting, etc. it might be useful to find user discussions related to such events from social media. Finding event related hashtags can be useful for this purpose. In this paper, we focus on the problem of… Show more

Users in social media often participate in discussions regarding different events happening in the physical world (e.g., concerts, conferences, festivals) by posting messages, replying to or forwarding messages related to such events. In various applications like event recommendation, event reporting, etc. it might be useful to find user discussions related to such events from social media. Finding event related hashtags can be useful for this purpose. In this paper, we focus on the problem of finding relevant hashtags for a given event. Features are defined to identify the event related hashtags. We specifically look for features that use similarities of the hashtags with the event metadata attributes. A learning to rank algorithm is applied to learn the importance weights of the features towards the task of predicting the relevance of a hashtag to the given event. We experimented on events from four different categories (namely, Award ceremonies, E-commerce events, Festivals, and Product launches). Experimental results show that our method significantly outperforms the baseline methods. Show less

Retrieving relevant tweets corresponding to cultural events
can be used in various applications like event reporting, event recommendation,
etc. This type of retrieval is challenging due to short length of the
tweet, noise, out of vocabulary words, abbreviations in the tweet. In this
paper, we focus on the problem of retrieving relevant tweets related to
given cultural event of a festival. We consider several factors like BM25,
DFR, presence of artist name, relevant hashtag… Show more

Retrieving relevant tweets corresponding to cultural events
can be used in various applications like event reporting, event recommendation,
etc. This type of retrieval is challenging due to short length of the
tweet, noise, out of vocabulary words, abbreviations in the tweet. In this
paper, we focus on the problem of retrieving relevant tweets related to
given cultural event of a festival. We consider several factors like BM25,
DFR, presence of artist name, relevant hashtag, festival name for finding
the relevance of tweets to the event. We apply BM25 + DFR model to
retrieve candidate set of tweets related to each event of a festival. We
find the top hashtags for each event by exploring meta-attributes of an
event. We re-rank the initial rank list from BM25 + DFR based on two
strategies, namely, presence of the event meta-attributes (artist name,
festival name, title, etc.) and the identified top hashtags in the tweet,
and based on the timestamp of the event. We experimented on a subset
of CLEF 2017 cultural microblog contextualization dataset. The experimental
results show that the proposed method is able to put relevant
tweets at the top of the retrieval list. Show less

In this paper, we describe a method to predict emotion intensity in tweets. Our approach is an ensemble of three regression methods. The first method uses contentbased features (hashtags, emoticons, elongated words, etc.). The second method considers word n-grams and character ngrams
for training. The final method uses lexicons, word embeddings, word ngrams, character n-grams for training the model. An ensemble of these three methods gives better performance than individual
methods. We… Show more

In this paper, we describe a method to predict emotion intensity in tweets. Our approach is an ensemble of three regression methods. The first method uses contentbased features (hashtags, emoticons, elongated words, etc.). The second method considers word n-grams and character ngrams
for training. The final method uses lexicons, word embeddings, word ngrams, character n-grams for training the model. An ensemble of these three methods gives better performance than individual
methods. We applied our method on WASSA emotion dataset. Achieved results are as follows: average Pearson correlation is 0.706, average Spearman correlation is 0.696, average Pearson correlation for gold scores in range 0.5 to 1 is 0.539, and average Spearman correlation for gold scores in range 0.5 to 1 is 0.514. Show less