Research Of Chinese Books Contents Automatic Analysis

As an important method of improving modern agriculture, agricultural books plays an important role in enhancing agricultural science, promoting agricultural research and improving agricultural workers’ quality. However, research on inner informational organization and knowledge structure of agricultural books has been lagged behind so far. Therefore, relative information and knowledge content of agricultural books can not be manifested, which lower knowledge content degree of agricultural books and a serious impediment to the speed and efficiency of the agricultural workers use agricultural books.With viewing the agricultural book contents as a breakthrough point, author survey agricultural books contents deeply and explicit the agricultural books contents rule, then combined with computer technology, conclude an accurate and practical method of agricultural books contents automatic analysis. The aim of this method is to extract the contents and key points of agricultural books accurately and practically. This method can raise the speed and accuracy of people using agricultural books and improve agricultural resources utilization. It also has a positive effect on refining agricultural books research and promoting the combination of informatics and other subjects.The thesis combines qualitative research method with quantitative research method to illuminate the possibility of agricultural books contents automatic analysis by the help of sampling survey, coding technology, decision tree algorithm and grounded theory; then with the guide of sampling survey method, choose1009agricultural books and record research items such as the years, first titles, second titles and so on; Then, based on grounded theory, author use open coding, axial coding, selective coding to analyze first titles, second titles and third titles and conclude agricultural books contents classification rule; Next, under the guide of coding technology and decision tree algorithm, author build agricultural books contents automatic analysis by using that rule; at last, choose500agricultural books from provincial library to test the precision of the model and evaluate the model.