Could Baidu Index Forecast The Trend Of Stock?

As we all know, the Internet has now become the main or even the primary source for people to obtain data. As for the Internet this very large collection of information, the search engine is a fast and convenient channel, so that you could quickly find the precise information you need.According to the 37th "Statistical Report on Internet Development in China" China Internet Network Information Center CNNIC in January 29,2016 release, as of December 2015, the number of the Chinese netizens had reached 688 million, of which mobile phone users has reached the scale 620 million. Internet search engines as the information needs of search volume data provided by intermediary parties, search engines and Internet records of information resources can record more fully the extent of fluctuations in demand users for the specific information. Based on such a large number of Internet users, if the search engine recorded certain things or keyword search volume there is a substantial increase, it is likely to say that even these are predictive search keywords or things will become the next hot spot. In response, many researches show that the Internet search data corresponding to real-life events are highly correlated. For example:search aggressively movies, games, or popular music concern, the unemployment rate for influenza and other epidemic or reported trading volume on the Shanghai Composite Index, etc., are related to the real world have this close contact.Among them,2009 Ginberg, who use search data to predict trends in infectious disease epidemic, they found that the use of the flu-related search data can be effectively predicted flu outbreak, which compared to the CDC (Centers for Disease Control and Prevention) are forecasting can do two weeks in advance to know the results.For financial markets, the market price can be seen as the concentrated expression which is the all investors'expectation of their collection of all the relevant information. Therefore, the behavior of investors' financial information collection and the price fluctuations in the financial markets have a very good correlation. According to previous studies, searching on Google and the amount of research in financial markets related keywords found changes in market volatility and search keywords are showing a certain amount of financial ties. As Vozlyublennaia in 2014 using data from the Google search as a measure of the probability of a proxy for investor attention, the study found that investors concerned about the degree of increase in short-term stock index returns will also change dramatically.The research combining with search engine keyword search volume and the stock market, especially about Chinese stock markets is very few in domestic or in foreign. On the basis of the reference for the previous search behavior research, screening the keywords which is most relevant to Chinese stock market, and gather its Baidu search volume index. At the same time, I use the semantic classification to get the similar semantic keywords assigned to a group and use the one words can reflects their common characteristic best named it.Then through a certain method of search volume keywords construct a new index to guide future investment behavior. In this paper, my research period is from 1,2011 to January 8,2016,because I am just able to get the data of Baidu Index in this period. In addition to the Shanghai Stock Index study, this paper also study about the effect of which use the above method to invest the relevant industry index.After, Baidu keyword combinations predictive index for the Shanghai Composite Index, and the corresponding sector index on research, pick out the combination which can significantly predict the stock market as the later research subject, examine if making the above methods as a trading strategy is useful. The next article will therefore more than elected investment targets for the Shanghai index keyword combinations, as well as investment targets for the corresponding sector index on the Shanghai index performance in combination with the same period of comparison, and using Fama-French three-factor model to observe whether with respect to the Shanghai Composite index over the same period there was significant excess returns and whether it is an effective trading strategy.Based on the above results can be found that using the method introduced above, we can predict the Shanghai Securties Composite Index and the related Industry Index. While building suitable Baidu keyword combinations not only a good predictor of the Shanghai index trend, at the same time it can be used as a very good trading strategies, with a huge excess returns. In the study also found that Baidu keyword index for predicting the effect of a combination of industry index is not as ideal as predicting the effect of the market, contrary to normal perception of people. The effect of prediction in a bull market is far inferior to it in a bear market, but due to limit the amount of available data, research time and research costs, the paper can not do more in-depth study.