Yuntian Shen analyzed the structure of the Chinese stock market compared to the US market and determined to use sentiments by individuals on Xueqiu.com as the source of data. Yuntian extracted sentiment data from stock social medias, including Weibo and Xueqiu, with web crawlers lively. He preprocessed Chinese text data with Jieba package and built an LSTM model to classify positive and negative sentiments. Our team utilized the proportion of positive sentiments to build a market prediction and trend analysis model.

Yuntian Shen analyzed past food delivery employees’ salary data, built a NN model that calculated proper salary subsidy based on past data, and improved the front-end of the food delivery app.

Yuntian Shen worked as a ML engineer assistant at ML and Database Joint Lab. He collected data of products of fund companies in current market and predicted their future profit with NN to analyze the products of each company. He also studied abnormal transaction data and violation detection and assisted to program quantitative investment algorithms.