Research

Research Interest

• Artificial Intelligence
• Machine Learning
• HCI
• Natural Language Processing

Research Experience:

Sentiment Analysis of Reader’s Comments in ESPNcricinfo: An Empirical Study (2016-17) Undergraduate Thesis (Github Link)

Supervised by: Dr. Anindya Iqbal

Sports news websites, like ESPNCricinfo, has become very popular nowadays, through which users can easily get updated about ongoing games, and express their opinions and enthusiasm about sports. The massive amount of user comments is a useful source that carries mass sentiment and opinions towards the game. There is no previous research on classifying sentiment of user’s comment in sports domain especially on ESPNCricinfo which is based on the game of cricket. In this work, we use a sentiment analysis algorithm and a lexical analysis tool to investigate the influence of various factors such as players form, type of the game,or even the time of the day on sentiments in the context of ESPNCricinfo. We start our analysis by looking at direct correlations, e.g.,we observe more positive sentiments on during the good form period of a player,weekends,more sentiment on the 4th day of a test match than others day.We then extend this basic analysis by investigating how team combination(visitor, host)pair affect the sentiment present in the comment.We also show that user’s sentiment differ for the same pair of team in different format of the game. The main contribution of this paper is the idea of using sentiment analysis to identify the major factors influencing user sentiments because we believe that understanding factors influencing the mood of users has interesting applications in advertising, recommendation, and search engine optimization.

Chakriguru: A job portal for illiterate people (2017-18)

Mentor: Syed Ishtiaque Ahmed

In urban area, mainly economically challenged people are those who have little or no education and hardly have any access to existing technologies. Apart from their physical labor they have little to offer as their skill levels are typically very low. So, often it is very hard to find appropriate jobs for them. On top of that whatever job they can get may not be safe or secured. And as they cant easily access mobile technologies (e.g using smartphone) they are basically detached from the huge opportunity that internet today can provide. We tried to find a way to break this technology barrier. We interviewed around 35 people from different occupations and classes to find out what they expect from a job portal. We gathered a lot of insights from their experience and tried to build a noble system which can serve these people properly. We participated in the Brac Urban Urban Innovation Challenge with this project and was able to reach the semifinal of the competition.

You can get more information about this project from our website http://chakriguru.com/ and https://drive.google.com/open?id=0B-mnPn6HGjfRWllLN1BUT3Vtb1k

Screen Reader for Visually Impaired People

The goal of this project is to build a screen reader application in android operating system using custom TTS Engine. Currently available screen reader application doesn't give full assistance for visually impaired people. The system can read out screen contents of the mobile device by producing voice sounds to describe menus, icons, dialogue boxes, files, and folders, task bar, etc. of the system in Bangla language. Where texts of multiple-languages (English and Bangla) are present, the system is able to read out English along with Bangla texts.