An Experiment On Tracking Stock Index

China’s stock market has been moved gradually to mature with rapid development. With the well-established institutions and widely-used computer technology, the kinds of stock index and creating index methods are constantly improved. Stock index has been played a significant role in stock investment and an increasing number of investors do their invest based on certain stock index.This paper adopts Sparse Principal Component Analysis (SPCA). Firstly, select certain stock as sample from Shanghai Composite Index (SCI) and Shenzhen Composite Index; secondly, find solution based on Minimizing Tracking-error Stock Prediction Model; thirdly, obtain stock forecasting combination. In this research, quadratic programming is applied to tracking stock index. The experiment result shows that:1. The stock combination selected by the method of SPCA has the advantages of interpretability.2. The stock combination within sample also has good performance outside sample.3. The Minimizing Tacking-error Stock Prediction Model makes it possible for more profit on less stock number, offering new investment idea for individual stockholders.