Application Of Double-layer Stock Product Recommendation Algorithm Based On Collaborative Filtering

In the traditional stock market research,the classical mathematical statistical model is often used as the representative to establish the stock prediction model and recommend the stock portfolio with good stock price prediction effect to investors.However,with the increasing diversification of investors’ needs,the problem that the previous traditional model did not fully consider the investors’ personal investment preference has become increasingly prominent.In view of this,considering that the collaborative filtering recommendation algorithm will not only fully consider the preferences of individual investors,but also have a relatively mature theoretical system,this paper integrates the collaborative filtering algorithm into the recommendation of stock products,which is similar to recommending items of interest to customers.In this paper,a single stock is regarded as a single item,and the stocks recommended to investors are regarded as items recommended to customers,The selected characteristic index value describing the stock is considered as the historical score of the customer for the item,so as to use the collaborative filtering algorithm.In order to make the recommended stock portfolio have higher returns than the original stock portfolio,this paper considers to analyze the A shares in some industries by using the double-layer recommendation model combined with collaborative filtering and multi factor stock selection.That is,first use the multi factor stock selection method to select some high-quality stock sets from various industries;At the same time,according to the historical investment stock names of investors,the nearest neighbor stock set of single-layer recommendation model is generated through collaborative filtering recommendation;Then the intersection of the nearest neighbor stock set and the high-quality stock set is taken as the result of the double-layer recommendation model.From the model recommendation results,it can be found that the best recommendation combination in this paper is top＿5.The recommendation results of the double-layer model have a good income effect in some industries;Comparing the single-layer and double-layer recommendation models,the double-layer recommendation model can optimize the single-layer recommendation model to a certain extent.In this paper,this model is called the double-layer recommendation model based on collaborative filtering.The theoretical framework is easy to understand.Especially for ordinary individual investors,it does not need a large number of investors’ personal privacy information,and has good operability and practical value.