The Research Of Market Forecasting Mode And Application Based On Network Search Index

The timely and effective market forecasting has always been the focusing issue of theory and practice. Due to the limitation of traditional forecasting technology, it restricted the effectiveness of prediction and expansion of application field. In recent years, with the development of search engine technology and change of people’s channel of obtaining information, thus, the brand-new prediction technology based on network search data emerged. However, existing researches are mainly at initial stage and concentrated on overseas, and most data source of network search originates from Google; meanwhile, it has not reached an agreement in the selection of search keyword, network search index synthesis and setting of forecasting model, and its practical application effect in different fields remains to be further researched.Based on the above background, as a starting point, this article explored the predictive value of the domestic network search data. First of all, this article established the theoretical analysis framework on consumer correlation between network information search and product sales. And on this basis, the system of network search index synthesis had been built. Then, by constructing analysis model, it selected instance object to carry out empirical theory. Finally, it surveyed the comparative study of the effect of the prediction model and comparative analysis of the network search index source difference. The major research contents and conclusions are as follows:1. This paper surveyed the value of market forecasting on domestic network search data. The empirical research result exampling automobile in this article showed that, although there is certain crossing relation between network search data the information contained by product historical sales data, it also included too much new information that is excluded from historical sales data, which can obviously promote the explanation effect of forecasting model to fluctuation of future product sales volume. Meanwhile, the new information contained by network search data is more sensitive to recent fluctuation of product sales volume with strong forecasting timeliness.2. It established the market forecasting model based on network search index and verified forecasting effect. The empirical result showed that, to compared with the contrast model, the forecasting model of product sales volume established by this article not only has high forecasting precision, but also shows strong timeliness. The forecasting of the monthly sales volume of subdivided vehicle model Passat indicated that, the degree of fitting was over 81.8%, the forecasting mean absolute error was 0.122, and the data was at least one month in advance of that issued by traditional statistical department. Particularly, in the aspect of introduction of the data of forecasting model network search, the improvement of the synthesis method of network search index made by this article proofed to be simple and effective, which, compared with the search data processing methods of time difference correlation synthesis and gradual synthesis, has better effect.3. It discussed about the difference of forecasting effect of different network search index sources. This article compared and surveyed the model forecasting effect by introducing Baidu indexes and 360 indexes. The result showed that, there was obvious correlation between the network search data and product market sales volume in these two index platforms. Generally, the explanation of the forecasting model based on Baidu index source was stronger(R-squared=0.818>0.711). The preliminary analysis of the reason causing this difference indicated that, there were various differences between these two network search index sources.The beneficial expand of difference discussion between synthesis index of network search and network search data source made in this article enriched the relevant theoretical research based on network search index; meanwhile, the empirical analysis of automobile industry showed that the market forecasting model established by this article has strong practical forecasting effect, which can provide effective decision reference for similar product market.