Analyzing Time Series Changes of Correlation between Market Share and Concerns on Companies measured through Search Engine Suggests

Analyzing Time Series Changes of Correlation between Market Share and Concerns on Companies measured through Search Engine Suggests Takakazu Imada1, Yusuke Inoue1, Lei Chen1, Syunya Doi1, Tian Nie1, Chen Zhao1, Takehito Utsuro2, Yasuhide Kawada3 1Graduate School of Systems and Information Engineering, University of Tsukuba 2Faculty of Engineering, Information and Systems, University of Tsukuba 3Logworks Co., Ltd. Abstract This paper proposes how to utilize a search engine in order to predict market shares. We propose to compare rates of concerns of those who search for Web pages among several companies which supply products, given a specific products domain. We measure concerns of those who search for Web pages through search engine suggests. Then, we analyze whether rates of concerns of those who search for Web pages have certain correlation with actual market share. We show that those statistics have certain correlations. We finally propose how to predict the market share of a specific product genre based on the rates of concerns of those who search for Web pages. Keywords: search engine suggest, topic model, market share, products genre, aggregation 1. Introduction Search engine companies recently examine how to predict specific information in real world such as economic trend 1, results of elections2, spread of influenza3 4 by analyzing concerns of those who search for Web pages within the search engine log data. For example, in the analysis of correlation between economic trend and search queries re- ported by Yahoo! JAPAN Big Data Report5, search queries that have high correlation with economic boom as well as economic depression are identified. Thus, it has become popular recently to analyze concerns of those who search for Web pages and to compare them with people’s behavior in real world. Considering those recent studies on correlations between concerns of those who search for Web pages and real world statistics, this paper proposes to utilize a search engine as a social sensor which is to be used for predicting market shares. In this paper, we measure concerns of those who search for Web pages through search engine suggests. Let us consider a case where we are given a query keyword “パ ナソニック” (Panasonic). Here, the search engine collects user search logs including the query keyword “パナソニッ ク” (Panasonic) and then presents suggests keywords such as “洗濯機” (washing machine), “ブルーレイ” (Blu-ray), “ 冷蔵庫” (refrigerator), and “ビエラ” (VIERA) which have strong relation to the query term “パナソニック” (Pana- sonic). In our framework, given the query company names for an- alyzing concerns of those who search for Web pages, we first collect search engine suggests of those query company 1http://docs.yahoo.co.jp/info/bigdata/ economiccondition/2014/01/ (in Japanese) 2http://docs.yahoo.co.jp/info/bigdata/ election/2014/02/ (in Japanese) 3http://docs.yahoo.co.jp/info/bigdata/ influenza/2014/01/ (in Japanese) 4http://google.org/flutrends/about/how. html 5http://docs.yahoo.co.jp/info/bigdata/ (in Japanese) names. Next, by specifying those query company names as well as search engine suggests, we retrieve Web pages for all of the search engine suggests and collect them into a mixture of Web pages for several companies that are com- petitive with each other in certain product genres. Then, we apply a topic model …

Content truncated to 500 words