Analysis on the Correlation of Traffic Flow in Hainan Province Based on Baidu Search

Internet search data records user’s search attention and consumer demand, providing necessary database for the Hainan traffic flow model. Based on Baidu Index, with Hainan traffic flow as example, this paper conduct both qualitative and quantitative analysis on the relationship between search keyword from Baidu Index and actual Hainan tourist traffic flow, and build multiple regression model by SPSS.


Introduction
Baidu Index is a data sharing platform based on huge amount of Baidu users' behaviour data. Through Baidu Index, we can acquire the actual search volume of relevant keyword with certain Baidu Index value. We build the consumer tourist behavior model based on the travel destination selection process, then search the primary keyword from the model by Baidu Index recommendation function to acquire corresponding keyword search volume. Finally, we build multiple regression model to analyze the relationship between the search index and the historical tourist flow of Hainan from 2014 to 2016 provided by Hainan Bureau of Statistics, and draw the conclusion.

Keyword selection
According to the relevant analysis of the degree of concern to the tourist destination before the tourists travel, the key points related to the destination (Hainan) traffic flow are selected according to the name of the tourist destination (Hainan), the actual route of travel, the destination of tourism destination and other tourist destinations word. According to this, this article selected "Hainan", "Hainan ticket(HNTI)", "Sanya ticket(SYTI)", "Hainan high-speed rail(HNRH)", "Hainan car rental(HNCR)", "Hainan traffic" as the benchmark key words, and then these benchmark keywords Baidu index search, Step to find out the actual search volume of these keywords, the second step to find out the relevant keywords, the third step, the use of webmaster tools such as the test, screening out the amount of search accounted for the first 80% of the key words: "Hainan high-speed rail", "Hainan ticket (Haikou + Sanya)", "Hainan car rental" 3 key words. According to this, the influence of the search volume of these three keywords on the traffic flow in Hainan is revealed, and the influence of the network search volume on the actual traffic flow in Hainan is revealed. Obviously, the actual search volume of Baidu index related keywords and Hainan tourism tourists the actual throughput of the trend is basically the same, with relevance.

Analysis of Traffic Flow Correlation
The first step, the data for descriptive analysis, mainly to determine whether the data collected is stable (Table 1). The second step, the data summary, mainly to determine the degree of fitting the model ( Table 2). The third step, the variables of multiple regression analysis, analysis of variance (Table 3). The F statistic of the model is 96.097 and the Sig value is 0.000, which is less than 0.05, indicating that the model to be established is statistically significant.

Establish traffic flow forecasting model
Through the analysis of a large number of the above data, and through the application of SPSS software, the establishment ( Table 4,Table 5): Among them, Y on behalf of Hainan tourism tourists forecast throughput, GTHN on behalf of Baidu Index Key words "high-speed rail Hainan", JPHN on behalf of Baidu Index Keyword "ticket Hainan", ZCHN on behalf of Baidu index keyword "car rental Hainan".
In the "Hainan high-speed rail", "Hainan ticket" and "Hainan car rental" through the significance test, at 5% significance level, I believe that "Hainan high-speed rail", "Hainan ticket" and "Hainan car" Can be through the statistical T test, so that "Hainan high-speed rail", "Hainan ticket" and "Hainan car rental" on the Hainan passenger throughput there is a significant impact.
From the collinear diagnosis table, we can see that the tolerance of "Hainan high iron", "Hainan ticket" and "Hainan car rental" is more than 0.1, and the variance expansion factor is less than 5, so there can be no multi-collinearity. From the above results, we can see that when the other conditions of the same conditions, the "high-speed rail Hainan" indicators for each unit of growth, Hainan passenger throughput increased by 2.086 units; "ticket Hainan" indicators for each unit, Hainan passenger throughput growth 0.984 units; "Hainan rental car" indicators for each unit of growth, Hainan passenger throughput increased by 7.14 units.

Analysis conclusion
Using tourist flow of Hainan as the object, based on the actual tourist flow of Hainan and the Baidu index search data of keyword including "Hainan high-speed rail", "Hainan airline ticket", "Hainan car rental", this paper analyze the relationship between the tourist flow of Hainan and Baidu keyword with econometrics theory, and reach the conclusion through the designed model. Based on the results of the above regression analysis, we can draw the following conclusions: (1) "Hainan high-speed railway" Baidu Index search volume and Hainan tourist flow presented a significant positive correlation, with correlation coefficient of 2.086. We can assume that "Hainan high-speed" Baidu Index search volume drive Hainan tourist flow to move positively. The higher "Hainan high-speed" Baidu Index search volume is, the higher is the Hainan tourist flow. On the contrary, the lower "Hainan high-speed" Baidu Index search volume is, the lower is the Hainan tourist flow.
(2) "Hainan airline ticket" Baidu Index search volume and Hainan tourist flow presented a significant positive correlation, with correlation coefficient of 0.984. We can assume that "Hainan airline ticket" Baidu Index search volume drive Hainan tourist flow to move positively. The higher "Hainan airline ticket" Baidu Index search volume is, the higher is the Hainan tourist flow. On the contrary, the lower "Hainan airline ticket" Baidu Index search volume is, the lower is the Hainan tourist flow.
(3) "Hainan car rental" Baidu Index search volume and Hainan tourist flow presented a significant positive correlation, with correlation coefficient of 2.086. We can assume that "Hainan car rental" Baidu Index search volume drive Hainan tourist flow to move positively. The higher "Hainan car rental" Baidu Index search volume is, the higher is the Hainan tourist flow. On the contrary, the lower "Hainan car rental" Baidu Index search volume is, the lower is the Hainan tourist flow.