Commodity Price Evaluation Based on Improved Data Mining Methods

Commodity price forecasting has become a key research point for market economy because it is very important for the development of industry, agriculture and finance. Commodity prices are affected by more and more market factors, resulting in unstable rules of change, which places high requirements on the prediction model. This paper uses the process of web data mining to establish a KNN model for website product information, and improves the distance and parameter impact on the original KNN algorithm to make it better fit the data set in this article, the accurate rate increased by 1.87%. KNN and improved KNN algorithm appropriately classify the price of website products to determine the current price and the current price of website products on the market. On the basis of improved KNN algorithm classification, the predicted price of the commodity is obtained by using decision tree regression, which is better than that obtained by direct decision tree regression. The performance of the proposed model is verified by testing the second-hand notebook data information crawled on the JD platform.