Search

The Tamil language is spoken by 80 million people around the world. The translation between Tamil and English leads to a significant impact by helping in the understanding of Tamil scripts, which ot... more...

Fake news detection is a significant problem where information is available from multiple sources across the internet. Most of the research on fake news has only targeted politics-related articles, but such models would not be robust enough to tackle fake news in the real world. To solve this problem, this research work incorporated transfer learning using attention-based transformers (BERT, RoBERTa, XLNet, DeBERTa, GPT2) and trained them on multi-domain datasets FakeNews AMT and Celebrity across different domains i.e. Politics, Entertainment, Sports, Business, Education and Technology. The proposed model has obtained state-of-the-art results while doing multi-domain and cross-domain testing, having beaten previous papers conformably. Also, the model has achieved a 99.3% accuracy on FakeNewsAMT and 84% accuracy on celebrity dataset. We believe the synergy of transfer learning in a multi-domain setting will make a robust model, which would be relevant in the real world. This idea originated from the fact that multi-domain research’s critical challenge is that data distribution is varying, and the key benefit of transfer learning is that it can perform well even when it is trained and tested on different data distributions. more...

In recent years, there has been a massive increase in multimedia data due to the increasing use of social media and communication technology. So, extracting useful information from a large set of data has become difficult and very time consuming. Multimedia Data (Text, Image, Audio, Video) summarization is a very useful technology that overcomes this challenge by eliminating information that is redundant or useless, and extracting only the relevant key details of the events in summaries. A lot of work has been done in this field to generate summaries in the form of text and images, but very limited research has been done to produce a multimodular summary especially on Asynchronous Data. This research work proposes an ILP based model, which takes a multimodal dataset (text, images, videos) as input and generates textual and image video summary as output. The results were obtained by comparing the two basic baseline’s ROUGE values with the proposed model. Results for different modalities have confirmed that the proposed model performs better than the other baseline approaches. more...

The purpose of this study is to conduct a scientometric inspection for the research based on WordNet lexicon. WordNet is a lexical database for the English Language developed by George A. Miller and his team at Princeton University in 1985. This study reviews WordNet-based research mainly published in the Web of Science (WoS) database from 1995 to 2019. The publication data has been analyzed computationally to present year-wise publications, publications growth rate, country-wise research publications, authors who published a large number of papers in this field. The results have been shown in the form of figures. The present study will be useful for a better understanding of the trajectory of the research work done in the field of WordNet during the last 25 years and shall highlight the prominent research areas and categories in the field of WordNet. more...

Most of the diseases spread from human-to-human which makes it dangerous for anyone who comes in contact with an infected being. The key point to note here is that such a spread gives rise to a network where nodes represent humans and edges show if two humans came in contact with each other or not. Studying and analysing pandemic networks help in managing the spread of disease efficiently. The recent coronavirus (COVID-19) outbreak that was first identified in December 2019 in Wuhan, Hubei province, China according to WHO reports, is an apt example of a deadly contagious disease. India has been fighting the virus since February 2020 relentlessly, with the scientists experimenting to find a cure for the disease, and healthcare personnel along with other essential workers ensuring that all necessary preventive and protective measures are being taken to reduce health risks. In our proposed study, we aim at forecasting the spread of COVID-19 in India with the help of SEIR-DH and linear regression model, by simulating the dynamics of disease spreading in a large population. We also aim to mathematically depict how increasing the severity of social distancing can affect the spread of the disease. The results of our study indicate that increasing the strictness of social distancing measures can help reduce the overall number of infected patients and also help flatten the epidemic curve of COVID-19 spread. The curve depicts the number of infected patients requiring healthcare for combating the disease over time. more...

Word Embedding plays a crucial role in natural language processing, and other related domains. The vast variety of language modelling and feature learning techniques often concludes in a quandary. The motivation behind this work was to produce comparative analysis among these methods and finally use them to flag hate-speech on social media. The progress in these word embedding techniques has led to remarkable results by incorporating various natural language applications. Understanding the different context of polysemous words is one of the features that evolved over time with these word embedding models. A systematic review on varying word embedding methodologies has been performed in this paper. Various experimental metrics have been used and detailed analysis has been done on each word embedding model. It is shown that analysis involves various aspects of the model like dealing with multi-sense words, and rarely occurring words, etc., and finally a coherent analysis report is presented. The various models under analysis are—Word2Vec (Skip-Gram, CBOW), GloVe, Fast-Text and ELMo. These models are then put to a real-life application in the form of Hate Speech detection of twitter data, and their individual capacities and accuracies are compared. Through this paper we show how ELMo uses different word embeddings for polysemous words to capture the context. We show how Hate speech can be better detected by ELMo because such speech requires better understanding of context of words for segregation from normal speech/text. more...

Word embedding or universal embeddings are the representation of textual data into vectors of real numbers. It acts as a link between the human understanding of a text to that of the machine. It maps the high-dimensional textual data into a vector space of low dimension, which represents the complex relationships existing in the data. Thus, it boosts the performance of tasks involving natural language processing. To know about the popularity gain and the impact of this field, research literature between 2000 and 2019 is analyzed with the help of a scientometric mapping for research done in word embeddings. The paper visualizes year-wise analysis, demographic analysis, category-wise distribution, and document type-wise distribution of the publications indexed in Web of Science (WoS). The paper also comprises a comparative study of Word2Vec, FastText, global vector representation of words, and bidirectional encoder representations for transformers. Pre-trained models are used for experimental comparison. Performance is measured by calculating the deviation of the similarity score of two words given by the models from manually assigned similarity scores by experts, repeated over a list of words on various datasets. The least deviation is shown by FastText due to the usage of morphological information in the skip-gram model and n-gram architecture. more...

The exponentially evolving size of information today has made it difficult to find relevant information quickly and efficiently. A good extractive text summarizer not only provides the most significant information from the document but also helps the user to decide the relevance of the information. The proposed method is a knowledge-based, generic, extractive text summarization technique. Our approach is based on the centrality of a sentences in the graphical representation of the documents. The graph is constructed using the pair-wise softcosine similarity measures between the sentences derived using the S semantic relations presented in WordNet lexical database. Eigenvector centrality measure outperforms the weighted degree, betweenness and closeness centrality measures. The resultant summary is compared against the gold-standard summaries of BBC news articles from year 2004 to 2005 and DUC 2007 datasets. The ROUGE-I, -2 and -Lmetrices are used to evaluate the results and found that our approach performs better than LexRank, TextRank, Luhn and LSA baseline text summarizers. more...

Centrality measures have evolved over the years and used over a variety of networks. Being one of the most basic measures to identify important nodes in ever-increasing modern-day networks, it has been manipulated and modified in every way possible to fit the requirement of the network and the way important nodes are perceived in it. Different centrality measures which seem to perform quite differently on a theoretical basis, provide similar results when applied to real-life networks. The centrality measures are studied based on the approach used, application areas, performance and measure the correlation among 14 centrality measures across 12 network topologies using Pearson, Spearman and Kendall correlation coefficients. more...

Social media enables the sharing of opinions, ideas, interests, thoughts, hobbies, etc., by creating social platforms such as Twitter, Facebook, TripAdvisor, Yelp, Rooter, Goodreads, etc. The above sharing generates a lot of information over various social media platforms, which is used in different research areas. Among the many research areas, one is opinion mining, which identifies and extracts subjective information from source materials by using text analysis, natural language processing, and computational linguistics. The application of opinion mining is abundantly seen in the areas of marketing, governance, tourism industry, etc. However, the application of opinion mining is not much explored in the area of user acceptance testing. Therefore, in this research article, we propose a model for opinion mining and use it as a supporting tool for user acceptance testing. Specifically, we use the opinion mining on “Twitter” tweets to analyze the response of an Open Beta, i.e., Public Beta versions of the software to quantify acceptance criteria qualities. Hence, we try to inculcate opinion mining in earlier stages of software development to aid user acceptance testing in order to get the user feedback as accurately as possible to make change in the concerned software before its final version is released. The model is implemented and evaluated for Open Beta versions of IOS 13 and AndroidQ using machine learning techniques such as decision tree, K-nearest neighbor, multilayer perceptron, and logistic regression. Our evaluation shows that we achieved the best result with K-nearest neighbor having an accuracy of 97.12%, which is 14.37% higher as compared to the highest accuracy ever obtained in this area. more...

An early analysis of growth dynamics for infectious diseases, like COVID-19, is needed to dissect the crucial driving factors that result in rapid disease transmission, refine the measures taken to control the pandemic and improve disease forecast. The phenomenological models are used to identify the initial climbing growth period of COVID-19 outbreak in India and have modelled 3 major epidemic growth models: Generalized logistic growth, Logistic growth and Generalized growth, to predict the growth in the total number of positive cases, daily increase in the number of positive tested cases and the daily growth rate in confirmed positive cases, dated from Apr 10, 2020, to Apr 20, 2020. The bootstrap resampling method is applied for data prediction to process the sample data, dated from Jan 31, 2020, to Apr 10, 2020, and to calculate the 3 major growth parameters: r (Rate of growth at an early stage), K (Final epidemic size) and C(Number of aggregate cases at time t), which are used to calculate confidence intervals which predict the future direction of the curve and increase in the number of confirmed cases with 95% accuracy for the interval Apr 10, 2020, to Apr 20, 2020. Our models predict exponential and sub-exponential spread rate in the number of positive cases in India from Apr 10, 2020, to Apr 20, 2020. Our findings reveal that significant measures are needed to control the transmission rate of the virus in the community, as the models predict sub-exponential growth in India. more...

Social networks have become a critical part of our lives as they enable us to interact with a lot of people. These networks have become the main sources for creating, sharing and also extracting information regarding various subjects. But all this information may not be true and may contain a lot of unverified rumours that have the potential of spreading incorrect information to the masses, which may even lead to situations of widespread panic. Thus, it is of great importance to identify those nodes and edges that play a crucial role in a network in order to find the most influential sources of rumour spreading. Generally, the basic idea is to classify the nodes and edges in a network with the highest criticality. Most of the existing work regarding the same focuses on using simple centrality measures which focus on the individual contribution of a node in a network. Game-theoretic approaches such as Shapley Value (SV) algorithms suggest that individual marginal contribution should be measured for a given player as the weighted average marginal increase in the yield of any coalition that this player might join. For our experiment, we have played five SV-based games to find the top 10 most influential nodes on three network datasets (Enron, USAir97 and Les Misérables). We have compared our results to the ones obtained by using primitive centrality measures. Our results show that SVbased approach is better at understanding the marginal contribution, and therefore the actual influence, of each node to the entire network. more...

This research article deals with translating a sentence from one language toanother. Our motive is to provide a technique that converts web-pages from one language to another automatically using Resource Description Framework (RDF) graphs. We particularly focus on making translation independent of human interference. Therefore, this research paper presents a framework and provides its implementationfor converting English sentences to Hindi using RDF and Pythonprogramming language. As a result of our research work, we successfullytranslatedour sample sentences present in English to Hindi. more...

Keywords are a summarized shortened version of any document. While a lot of research has been done on keyword extraction, very few of them analyze the network of a semantic network to identify the most important words in the document. In this research, we present one such method which uses WordNet as our knowledge base to exploit the semantic relatedness of terms and hence determine keywords. This is based upon graph centrality measures which help to identify the central nodes or keywords from the document. more...

Sentiment analysis research of public information from social networking sites has been increasing immensely in recent years. Data available at social networking sites is one of the most effective and accurate source to identify the public sentiment of any product/service. In this paper, we propose a novel localized opinion mining model based on common sense information extracted from ConceptNet ontology. The proposed methodology allows interpretation and utilization of data extracted from social media site “Twitter” to identify public opinions. This paper includes location specific, male- female specific and concept specific popularities of product. All extracted concepts are used to calculate senti_score and to build a machine learning model that classifies the user opinions as positive or negative. more...

Hindi, the majority language of India, is still in its infancy stage concerning to natural language processing applications and research. Spelling detection and correction for Hindi language is an important process which has received inadequate attention. Today, there is lots of work available for English Spelling detection and correction but for Hindi, not much work has been done. Unlike English which has only two types of letters i.e. 21 consonants and 5 vowels, Hindi language in addition to 40 consonants, 10 vowels, consists of various types of symbols (ten vowel sign (matras), half letters & halant etc.). So the methods available for English cannot be directly applied for Hindi. Various documents, newspaper, novels, books, magazines, government notice etc. are typed in Hindi so there is need for development of spelling detection and correction tools for Hindi. In this paper author focused on detection and correction of spelling errors in Hindi language for non-word errors. more...