Research On Quantitative Investment Strategy Of New Energy Industry Based On Text Mining And AI Algorithms

With the increasing development and quantitative level of China’s financial market,the use of artificial intelligence(AI)algorithms for quantitative investment has gradually become a growing industry trend.At the same time,more and more investors tend to obtain information and express their views on the stock market online.Therefore,from the perspective of investor sentiment,combining text mining and AI algorithms to construct a quantitative investment strategy is a further exploration of the application of quantitative investment.Currently,in the context of economic transformation and dual carbon,the new energy industry has ushered in good development opportunities.To this end,this article selects stocks in the new energy industry as the research objects,and use text mining technology and AI algorithms to construct a new energy industry quantitative investment strategy that meets the investors’ expected interests and risk demands,thereby providing better investment ideas.The main research contents are as follows:1.Public opinion factors for the new energy industry are constructed based on text mining.The stock bar comments of 148 new energy industry on Oriental Wealth Network from January 2020 to October 2022 are crawled as text data.By establishing a Naive Bayes sentiment classification model,sentiment analysis is performed on the stock review data to construct public opinion factors for the new energy industry.2.Examining the effectiveness of public opinion factors and candidate factors.The validity of industry public opinion factors is examined through methods such as Information Coefficient(IC)value inspection method,layered return testing method,and factors related inspection.The test results show that the constructed new energy industry public opinion factors are effective.Meanwhile,the validity of 15 candidate factors is examined,and four effective factors are selected and added to the multi-factor stock selection model."3.Quantitative investment strategies are conducted based on AI algorithms.Three AI algorithms,namely support vector machines,random forests and XGBoost,are introduced into the multi-factor stock selection model.And with the help of the AT-edu quantitative platform,rolling backtesting of the quantitative strategies is conducted.The results show that the returns of the multi-factor quantitative stock selection model constructed by the AI algorithms all outperform the returns of the CSI 300 benchmark in the same period,and the backtesting performance of the model added with the public opinion factor is even better.By comparing the performance indicators of the three AI algorithms stock selection strategies,it is found that the XGBoost stock selection model performs best.4.Conduct strategy optimization and validation,and provide reasonable and effective investment recommendations.By adjusting the position quantity parameters for strategy optimization,the results show that the combination of the top 10 ranked stocks has the best performance,with a cumulative return rate exceeding 100% and a Sharpe ratio exceeding 1.In summary,more accurate analysis and prediction can be achieved through text mining and AI algorithms,which helps investors to formulate more accurate investment strategies.This is of great significance for the innovation and breakthrough of quantitative investment strategies in the emerging new energy industry.