Agricultural Information Level Evaluation and Prediction Research based on Supporting Vector Regression

ABSTRACT

There are the following problems on research of the agricultural information level evaluation and prediction in China and other countries. (1) People do not find a generally accepted evaluation index system for the agricultural information. (2) The methods to determine the index weight are usually based on subjective judgment, cannot objectively reflect the correlation of index. (3)Most evaluation and prediction methods of agricultural information level study on social information from the economic viewing point and are based on linear transformation, they cannot exactly reflect the characteristics of nonlinear fitting between agricultural information index system and evaluation results. In order to solve the above mentioned problems, this paper first constructs a set of agricultural information index system, then gets the weight of each index according to the method of entropy and does evaluation and prediction of the agriculture information level with the support vector regression method. This paper, based on objective view and nonlinear method, is feasible, effective, has the value of application and popularization proven by empirical research.

Keywords: