Eric YIN

Master's

Management Information Systems

GUOTAI JUNAN Securities

Quantitative Research Intern

Developed a two-stage factor mining model based on residual networks, where data images were constructed, and residual networks were utilized to extract cross-sectional features, followed by a recurrent neural network (RNN) to extract time-series features. This model achieved a 14.76% return in a backtest on the CSI 300 index  Employed the earnings-to-price ratio (EP) as a single factor for monthly stock selection, applying outlier removal, standardization, and market capitalization as well as industry-neutralization processes to the factor. Calculated cumulative return curves for 10 stock portfolios, achieving an 8.7% excess return relative to the CSI 500 index.