Research Projects

iSchool Online Video Caption System

Syracuse University, June2015 – May2016

This project aims at establishing an online learning system with captioned video provided by students in crowdsourcing mode. Based on this system, we want to study if the length of the video would affect students’ editing accuracy.

Trust, Deception and Suspicion Prediction Based on Natural Language Processing
Syracuse University, July2015 – May2016

The main purpose of this project is to establish a prediction model for trust, deception and suspicion. We use a survey generated mostly by myself to choose the top 10 textual elements which affect people’s suspicion and trust most. Then we analyze the massive text-based conversation data and the trust, deception and suspicion level via natural language processing to build the model. Welcome taking our survey HERE!

Analysis on Different Classifiers for the Study of SMS Spam Filtering
Syracuse University, Jan2015 – Mar2015

The goal of this study was to test different classifiers on different vectors to get the best performance on SMS spam filtering.  I tested nine different classifiers on nine different vectors and set EM classifier as baseline. Results show that Multinomial Naive Bayes + Ngram(N=(1,3)) + Boolean has the best performance both on both training set and testing set.  [Course Paper]

User Behavior Analysis of “Showing Happiness” on Wechat Based on Social Comparison Theory
Beijing University of Posts&Telecommunications, Feb 2014 – June 2014

This thesis discussed the main factors affecting subjective well-being when showing happiness on Wechat. I constructed the research model based on the lateral comparison theory, introducing three variables: enhanced self-esteem, sense of personal worth and positive self-awareness to avoid negative feedback evaluation, to explore the relationship between these three variables and subjective well-being. By analyzing data from 235 surveys, I got the conclusion that positive self-awareness to avoid negative feedback evaluation don’t affect the subjective well-being directly but by affecting the other two variables that significantly affect the subjective well-being. [Bachelor Degree Thesis (Chinese)]

The Study on Blending Modes of Social Network and E-Commerce
Beijing University of Posts&Telecommunications, Dec 2012 – May 2013

This paper analyzed the blending modes of social network and E-commerce from two perspectives:social e-commerce and e-commercial social network. Corresponding suggestions were proposed as well.[Innovation Project Paper]