Somya Mohanty, Assistant Professor, Department of Computer Science
A Big Data approach towards understanding the factors responsible for scientific citations
Citations have been widely used as a measure of scientific impact. While this might be a controversial topic, the success of scientific research can be attributed to a wide range of factors, and it is necessary to understand the factors which contribute to citation success. The overarching goal of the proposed research is – to develop data-driven machine learning models which predict citation success, and also in turn identify the key features which influence citations.
In order to achieve this we aim to develop a generalized prediction model (for all research domains), while considering topic specific scientific impact factor into its analysis. Towards this, the research objectives are as follows:
- RO-1: To extract publication features and normalize citations across different domains.
- RO-2: To develop machine learning models for citation count prediction.
- RO-3: To evaluate the model-variable importance using factor analysis.
- RO-4: To understand the factor dissimilarity across domains.
Scope of CCI Lab Involvement:
The scope of CCI lab is to enable access to the Microsoft Academic Graph which will enable the study into the citation research.
This includes setting up a small storage allocation on Microsoft Azure.