Research And Application Of Data Analysis Algorithm Based On National Land Right Information Acquisition System

With the advent of the Big Data Age, the size and complexity of the data grows beyond the Moore's Law of hardware capacity growth, posing enormous challenges to machine processing and computing power, as well as providing access to vast amounts of data unprecedented opportunities.Therefore, there is an urgent need to study how to deal with the huge amount of data in-depth and effective analysis, find and extract the information hidden in them, and explore the relationships and rules existing in the data. The purpose of data analysis is to extract and concentrate the information in the massive data, so as to help people based on the conclusions drawn to develop the appropriate program.In the national land rights information acquisition system on-line and stable operation of the two years, accumulated a large number of data of confirmation and registration of land contractual management right.Including a total of 2778 counties (districts) quarterly information data of 9 quarters, each quarter quarterly information among counties (districts,cities) submitted quarterly data of not less than 42, as of the second quarter of 2016, the total amount of data is 1195837 records. In addition, there are 9 basic information data per year, the total amount of data for the 75961 records. In the process of system operation, we found that: the system in accordance with the quarter for data collection, the cycle is relatively long,and the number of data items collected is large, so the submission process is likely to occur inaccurate or incorrect data.Based on the above, how to use the method of data analysis, combined with the large amount of data already accumulated in the information collection system, establish the data forecasting model, provide the support for the land-based data decision quickly and timely, analyze and discover the potential inaccurate or incorrect data has become the focus of this paper,the main work and innovation are as follows:1. A data warehouse which can be used in the subsequent data analysis is designed for the original data structure of the Land Right Information Acquisition System, which is based on the core technologies involved in data analysis. The Structured Query Language script written for the land rights information collection system stored in the data warehouse automates the process of data extraction, transformation, cleaning and loading (ETL) from the source database to the destination data warehouse and can adjust the script as needed to provides granular control over data extraction rules, greatly improving the flexibility of data cleaning and loading.2. On a variety of data analysis algorithms are studied and compared,according to the overall design of the data warehouse and the nonlinear characteristics of the data, the scale transformation method of the input-output data is designed, the rules of the training sample set are selected, the training function and transfer function of the network is determined, the BP neural network hidden layer nodes and number of iterations is given,the BP neural network data model with three - layer structure is established on the basis of the key parameters above and the simulation are carried out under matlab environment. Under the condition that the data set is accurate,the normalized error of the prediction result can be used as the guideline data of the next quarter's data submission. Under the condition that the data exist inaccurate or incorrect can be found by the prediction with large deviation, which can further improve the accuracy of the data.