The Application Of Principal Component Analysis And Neural Network In Indus Trial Economy Data

Data mining, also known as knowledge discovery in databases, is currently ahot issue in the field of artificial intelligence and database research, it is a decisionsupport process, mainly based on artificial intelligence, machine learning, patternrecognition, statistical, database, visualization techniques and so on.“China Statistical Yearbook” is compiled by the National Bureau of Statistics ofPeople’s Republic of China, published by China Statistics Press, the book collectionof the People’s Republic of China and the system in various provinces, autonomousindividual, as well as municipalities in2012social statistics and all aspects ofeconomic data, the book also covers more than the previous year and the importantdata of the statistical data in various regions, which reflects the development ofsociety in all aspects of People’s Republic of China, is a statistical yearbook data.Algorithms used herein include principal component analysis algorithm andneural network algorithm. Principal component analysis is a linear transformationthrough multiple targets to reduce the dimension of the process of the few mainingredients, by a small number of principal components to reflect the data. Neuralnetwork is a neuron simulation data model established animal, the data modelstructure is similar to the brain information processing synapse coupling.This paper completed the following two tasks.The first task is through the use of principal component analysis to study the dataof2012industrial enterprises in a variety of indicators to China ’s31provinces,autonomous regions and municipalities, industrial enterprises develop acomprehensive analysis and evaluation of the18indicators of industrial enterprises dimensionality reduction to three principal components, principal component analysisalgorithm through principal component analysis to calculate the principal componentscores of31provinces, autonomous regions and municipalities ranked industrialdevelopment, analyze the various provinces, autonomous regions and municipalitiesof the features and advantages, provide theoretical basis for balanced economicdevelopment of our country.The second task is to establish a BP neural network prediction model, throughthe “China Statistical Yearbook “in the Industrial Products GNP data and data forresearch, build predictive models. Industrial Products index includes31indicators ofcrude oil, natural gas, salt, etc., there is a certain relationship between GDP andindustrial products, the paper through the relationship of GDP and industrial productsbetween the study of various indicators, BP neural network model to create, andcompared with the results of multiple linear regression, found that the effect on BPneural network is better.