Study On Tree Species Suitability Of National Reserve Forest Based On Data Mining

As one of the important natural resources and strategic resources of the country, forest land resources have a great responsibility for the survival and development of the forest, carrying the heavy responsibility of ecological construction, and have the dual mission of optimize environment and promote development. In order to solve the structural contradiction of the output of our country’s forest resources, the state builds the reserve base and practices the reserve forest system. The first principle of designating the reserve base is matching species with the site, so it is very important to evaluate the suitability of tree species. With the development of China’s forestry investigation system and the improvement of the Internet of things, forest resources data increasingly diversified and massive, but its utilization is not sufficient. The traditional research on the suitability of specific tree species mainly using statistical analysis method based on the sample investigation. But, it consumes a lot of manpower and material resources, and can’t use the existing data comprehensively. Also, it is difficult to deal with the complex nonlinear relationship between tree species suitability and environmental factors. Focusing on these problems, this paper integrated multisource forest resources subcompartment data, based on the data of subcompartment investigation, analyzed the suitability of forest tree species using the theory and method of data mining. Expected to discover species suitability knowledge from a large amount of data, to provide the assistant decision making and technical support for the construction of reserve forest base, and to provide a new idea and method for the suitability evaluation of tree species.The specific research contents of this paper are as follows:(1) Integrated multi-source data of forest resources in sub-compartments. This paper using forest resource investigation data,DEM data, the Soil Database, and meteorological data,to integrated a multi-source forest resources subcompartment data set oriented to the research of the suitability of forest tree species, by extraction, transformation, cleaning, integration and other steps. Expanding environmental factor information to provide a more comprehensive study on the suitability of tree species.(2) Comprehensive analysis of the technology system of forest tree species suitability data mining. This paper clarified the focus and objectives of reserve construction, analyzed the tree species suitability evaluation index and evaluation principle. Built a reserve forest tree species suitability evaluation index based on the existing data mining algorithms and general steps according to the actual situation of the data. Then put forward the process and the suitable algorithm for tree species suitability data mining.(3) Tree species suitability evaluation dimension reduction model. Many environmental factors related to the growth of trees. This paper analyzed the characteristics of forest resource subcompartment data, using rough set algorithm based on neighborhood granulation, chose out the important factors affecting the suitability of tree species.(4) Tree species suitability classification prediction model. According to the basic principle and calculation process of the BP artificial neural network algorithm and C5.0 decision tree algorithm, the modeling process and the concrete realization of the two algorithms applied to the prediction of tree species suitability classification are presented.(5) The experimental results and analysis. In this paper, a case study was carried out to evaluate the Pinus koraiensis Sieb. et Zucc. forest in Liaodong mountain area. A variety of experimental schemes are designed, the simulation accuracy of the traditional multiple linear regression model and the data mining models are compared. The rationality of the data mining models were analyzed and verified. By using the classification prediction models established in this paper, the suitability distribution of tree species with higher accuracy was obtained.The main innovations of this paper:(1) Introduce the data mining technology into the evaluation of the suitability of the forest reserves. Constructed the suitability evaluation index of the tree species based on the characteristics of the forest reserve. And put forward the complete technological process;(2) Using rough set algorithm combined with BP artificial neural network algorithm and C5.0 decision tree algorithm to achieve classification prediction, effectively improved the computational efficiency of the model, and improve the accuracy of data mining prediction.(3) Achieve the quantitative and qualitative evaluation of Pinus koraiensis Sieb. et Zucc. species suitability in Liaodong mountain area.