E-commerce Sites Search Results Relevance Prediction Based on Ensemble Approach
Qiqi Wang · 2017
Though there are numerous traditional models to measure the search relevance of search engine, the evaluation results of existing models are not precise enough and difficult in operation in most of the cases.This article proposed a forecasting approach based on ensemble method to improve the precision of search relevance prediction.The critical process of the approach requires features extraction and parameters selection.The experimental results show that it works well in given data set.Furthermore, in order to select the optimal weights set, we also give a model and algorithm for the problem in the end of the article.