Qualification: 
MSc, BSc
thusharamg@am.amrita.edu

Thushara M. G. currently serves as the Lecturer at the Department of Computer Science Applications, School of Engineering, Amrita Vishwa Vidyapeetham, Amritapuri. She has completed M. Sc. in Computer Science and currently pursuing her doctoral studies. She has worked as Research Scholar in Technical University of Munich, Germany (Jul 2010- Nov 2012). She had been invited to work in DALI Labs in University of Perpignan, France (Nov 2012 to Aug 2013).

Qualification: BSc (SS), MSc (CS)

Publications

Publication Type: Conference Proceedings

Year of Conference Title

2018

A. R Das, Afsal, P. M., and Thushara M. G., “Tagging of Research Publications based on Author and Year Extraction”, 2018 International Conference on Advances in Computing, Communications and Informatics (ICACCI). IEEE, Bangalore, India, 2018.[Abstract]


Tagging provides a way to give a token of link to research publications which facilitates recommendation and search process. This paper proposes an auto-tagging methodology using Conditional Random Fields (CRF) for the precise extraction of authors and publication year from the research publications. This proposed system use auto-tagging methodology for research publication recommendations. This mechanism suggests a research publication based on the corresponding tags they have created. The system also generates various dynamic reports for graphical analysis of projects carried out in each research domain. These statistics benefits help users to get an overview of the trends of research works done over the past few years and ongoing researches in each department. The proposed system helps the students, faculty and other academicians to get involved in ongoing researches and also to obtain ideas in their respective research domain. It makes aware staffs and students with new inclinations in research publications.

More »»

2017

Thushara M. G., SA, S., and S, S., “KEA based Document Tagging for Project Recommendation and Analysis”, ICACNI 2017 : 5th International Conference on Advanced Computing, Networking, and Informatics, At National Institute of Technology. Goa, 2017.[Abstract]


This paper proposes an innovative approach in managing project related documents, project domain analysis and recommendation of open areas from current project document pool. Using Keyterm Extraction technique, documents are tagged under appropriate categories and sub-categories for better management of project documents. Hence this tagged document serves as reference for the students who are planning to take up new projects. The system generates various reports for statistical analysis of projects carried out in each research domain. These statistics benefits users to get an overview of the trends of project works done over the past few years. There are also reports illustrating number of open areas over respective academic years. The open areas are identified and listed for the students. This novel approach would help the students who are seeking new project. Our system helps the students, faculty and other academicians to get involved in ongoing projects and also to obtain ideas in their respective research domain. We have modified the stemming method in basic Keyterm Extraction algorithm (KEA) by adding Porter stemmer rather than Lovins stemming method and our experimental results confirms that our modified Keyterm Extraction method outperforms the KEA method while tagging English documents.

More »»

2017

Thushara M. G., MS, K., and Nair, S. S., “Domain Classification and Tagging of Research Papers using Hybrid Keyphrase Extraction Method”, ICACNI 2017 : 5th International Conference on Advanced Computing, Networking, and Informatics, At National Institute of Technology. Goa, 2017.[Abstract]


Extracting thematic information from scientific papers has wide applications in information retrieval systems. Keywords give compact representation of a document. This paper proposes a document centered approach for automatic keyword extraction and domain classification of research articles. Here we induce a hybrid approach by adopting different methods in various phases of the system. Domain classification is important for researchers to identify the articles within their interest. The proposed system uses Rapid Automatic Keyword Extraction (RAKE) algorithm for automatic keyphrase extraction which gives best score of keywords. The classification process concerns semantic analysis which includes keyword-score matrix and cosine similarity. A comparative study of performance of RAKE algorithm which uses score-matrix against KEA algorithm based on term frequencies to extract relevant keyword was also performed.

More »»

2017

Thushara M. G., Krishnapriya, M. S., and Nair, S. S., “A model for auto-tagging of research papers based on keyphrase extraction methods”, 2017 International Conference on Advances in Computing, Communications and Informatics (ICACCI). IEEE, Udupi, India, 2017.[Abstract]


Tagging provides a convenient means to assign tokens of identification to research papers which facilitate recommendation, search and disposition process of research papers. This paper contributes a document centered approach for auto-tagging of research papers. The auto-tagging method mainly comprises of two processes:- classification and tag selection. The classification process involves automatic keyword extraction using Rapid Automatic Keyword Extraction (RAKE) algorithm which uses the keyword - score matrix. The generated top scored keywords are added to the train dataset dynamically, which can be used further. This add-on feature of the system is considered to be one of the main advantages of the system since adding new born phrases is time-consuming and error prone. Cosine similarity is used for classifying the research paper into corresponding domain utilizing the extracted keywords. Tag selection concentrates on the selection of proper tags for the research paper. Tagging facilitates better search facility and determines the dynamics of research areas in terms of number of publications in a domain by each author. The system generates reports for statistical analysis of research papers in each domain and expertise of each faculty in the research community.

More »»

2016

Thushara M. G., Jayaprakash, V., and A. Kumar, P., “Panel generation as an application of genetic algorithm”, Second International Conference on Computing Paradigms (International Journal of Control Theory and Applications), vol. 9, no. 10. International Science Press., pp. 4509-4518, 2016.[Abstract]


In this paper, a method which can generate academic project review panels by providing the Time Table (TT) of the faculties and a batch of students. The faculties and students in a college are allocated with a TT at the beginning of the academic year. The faculty TT contains the list of hours for which they have to engage a batch of students. The other hours are considered as their non-engaging hours. Likewise, in every final year student's TT there are fixed time slots for their projects. The area of concentration in this paper is to generate a review panel such that the non-engaged slots from the faculty's timetable are matching with the project hours of the batch so that, a review panel can be generated and allocated. The number of members in each project review panel is determined by the project coordinator. When the algorithm is used, the time, the effort, and the accuracy for generation of the review panel are far better than the one generated manually. © International Science Press.

More »»

2016

Thushara M. G. and Dominic, N., “A template based checking and automated tagging algorithm for project documents”, Second International Conference on Computing Paradigms (International Journal of Control Theory and Applications), vol. 9, no. 10. pp. 4537-4544, 2016.[Abstract]


Most project documents are designed based on templates. Some of the standard formats which are accepted worldwide are User Requirements Specification (URS), System Requirements Specification (SRS), System Test Cases (STC), User Acceptance Testing (UAT), Defect Track Log (DTL). This paper describes features and methods for document comparison with templates and score generation. The uploaded documents are categorized based on the research area. The approach described here exploits the results of recent progress in information extraction to compare a document with its template and calculates a score. Clustering the documents is an important task in applications that analyze the template structure of a document. Using keyword Extraction algorithms, documents are tagged under appropriate categories to enhance the accuracy of search results for the project documents. Thus the documents in the repository serves the purpose of reference for students who are searching for projects under a specific area. This research would also help the project mentors to check the template matching of the documents submitted by students. Also our system helps students who are seeking new project, faculty and other academicians to get involved in ongoing projects and also to obtain ideas in their respective research domain.

More »»

2015

R. Meenakshi, Jayalekshmi, G., Hariram, S., Shiju Sathyadevan, and Thushara M. G., “Visualization with Charting Library Based on SVG for Amrita Dynamic Dashboard”, Procedia Computer Science, vol. 58. pp. 371 - 379, 2015.[Abstract]


Data Visualization is the representation of data in a graphical or pictorial format. For the effective communication of data for a user proper visualization is necessary. Visualization is essential in order for the user to understand the data in an easy way. Visualization of data is done through various charts that represent the attributes of the data. For web applications, there are many open source JavaScript libraries that work on HTML5 (using SVG or CANVAS). But the drawback of these libraries is that they don’t provide for much flexibility with respect to configuration. They also don’t provide generalization of charts. Also many data mining algorithms are not supported by these libraries for data visualization. This paper has illustrated in building JavaScript charting libraries that would ensure proper visualization of data which is flexible for user customization. The charting library supports different types of charts varying from scatter chart to line chart to bar chart that are used for various algorithms. The libraries are built based on Object-Oriented JavaScript concept to support web applications that run either on the internet or intranet, so that extending the same in the future is also possible.

More »»

Publication Type: Journal Article

Year of Conference Title

2018

A. Das and Thushara M. G., “Auto-tagging of Research Publications using Keyword Extraction Technique”, Journal of Advanced Research in Dynamical and Control Systems, vol. 10, no. 06-Special Issue, pp. 1168-1172, 2018.[Abstract]


In recent years, there is a vast growth in the publications of research papers in various domain of research. The research articles need to be categorized into the corresponding domain. This categorization is a time-consuming and error-prone process. Using the keywords relevant to a particular domain, the grouping of these research papers can be done more effectively. That is each uploaded article is mapped to set of tags.

More »»

2018

Thushara M. G., Somasundaram, K., and Jayaraj Poroor, “Analysis of Numerical Accuracy in Floating Point Programs Using Abstract Interpretation”, AFMSS, Springer LNCS (in print), 2018.