基于网络舆情大数据预测商品销量

Predicting Commodity Sales Based on Big Data of Online Public Opinion

摘要: 本文对网络舆情分为5大类：从众舆论、客户评论舆论、经济舆论、社会舆论以及文化舆论，根据不同类型舆论选取其直接影响的商品。根据舆论微博指数确定舆论突发时间，再获取相同时间的商品淘宝指数，运用SPSS曲线回归，通过比较不同曲线的显著性确定舆论微博指数与商品淘宝指数间的关系及具体关系式。最后，我们得到88购物节舆论与女装、日用品、鞋、服饰配件四大类常见商品销量的关系。

Abstract: In this paper, we firstly divide the Internet Public Opinion into five categories: adrift public opinion, customer evaluation of public opinion, economic consensus, social public opinion and cultural opinion. Then, we screen out the commodities which are directly affected by the above five kinds of public opinions. According to the micro blog index of public opinions, we determine the burst time of public opinion and get the commercial taobao index at that time. By comparing the significance of different curves which is obtained by SPSS, we can confirm the relationship between themicroblog index of public opinions and commercial taobao index with the specific formula expression. Finally, we get the relationship between the public opinion of the 88 shopping festival and the sales of four categories of common goods: women’s clothing products, daily necessities, shoes and clothing accessories.