Research Area: Knowledge Graphs, NLP, Machine Learning
2021 - 2022
Extracting Cognitive Relationships
Between Citing and Cited Papers in Academic Research
In the past few decades, academic researchers have worked in different fields and shared their knowledge with us with the help of publishing articles, research papers, transactions, etc. Keeping track of such a vast library is getting difficult day by day. It is complicated, but it is also tedious and time-consuming. Many techniques have been invented to ease this job and keep this in mind; we are working differently to get even a vaster image of the previous track of those research papers or articles. Research papers may be represented as a networked information space with a collection of information items interconnected by direct connections, referred to as the Citation Graph. It is possible to augment the citation graph with meaningful relationships between the citing and cited articles to communicate the reason for the citation using meaningful semantic tags. We suggest extracting a cognitive relationship between citing and cited reports in academic research using the citation Context and Reasons Ontology (CCRO). It would be very beneficial to reasoning systems if these reasons were communicated formally using Ontology. We think that information present in the network of a research paper can be utilized by visualizing different information of research paper, i.e., Author Name, Publication time, etc. The main reason behind this step is to present vital details by detecting raw entities in academic research. We will create a knowledge graph of citing and cited publications and uncover the significance of individual papers. After careful formulation of a problem, we derived some steps from completing our tasks which are: (i) Extraction of Citation Markers (ii) Extraction of Citation Context (iii) Sentiment Analysis (iv) Finding Cognitive Relationship using CCRO (v) Selecting and using the proper Ontology (vi) Knowledge Graph Creation. Knowledge Graph, in this case, can be referred to as Citation Graph. Moreover, it can expose essential and intriguing information about the history of a particular academic research publication, including important and exciting events that occurred throughout the article's lifespan.
Research Area: Scientometric, Citation Analysis, NLP
2019 - 2020
Semantic Based Citation Reason Analysis using Reporting Verbs
Research is a continuous and recursive process. Every research paper and articles are built on some prior knowledge in the field. Research papers include citations to the external resources to discuss the work done by previous researcher. With the rapid development in research area, it becomes challenging for researches to recognize the quality research work. There are various existing approaches for the analysis of citations but most of these approaches rely on quantity rather than quality of the citations.
In the last few years, researchers questioned the methods of mere quantitative citations analysis, arguing that all citations are not equivalent and the reason for citation must be taken into consideration while counting. Lately, most of the classification methods have used various part-of-speech such as nouns, adjectives etc. as feature vector for citation analysis. This thesis proposed a Mapping Graph between opinion verbs used in citation texts and its use in various sentiments with the help of using Stanford Linguist Beth Levin's English verb classes. It also presents a novel technique for semantic based citation reason analysis by extracting and understanding the role of reporting verbs in citation texts with different sentiments. For experiments, we have used ACL Anthology Citation Dataset comprising of over 8,000 citation sentences for semantic based citation reason analysis.
Research Area: Multimedia, Parametric Identification, Z Indexing
2008 - 2009
Vector Shape Classification and Z indexing
Vector images are the compact representation of shapes using hard mathematics. Shapes are stored compactly
and render efficiently. The advantage of these vector images is that images can be fraction of size as compare to the bitmap. Much work has been done on digital images from different perspective not only to find detail parametric definition of shapes but also their display mechanism. However, this paper is to justify these shapes and then their order of display. Vector
shapes consist of different shapes such as line, ellipse, rectangle etc. It is necessary to identify these shapes and their mathematical formulas. We will discuss the shapes with their extracted parametrical identification using formulas and their drawl by using Z index.