Study On Rural Collective Construction Land Grading Based On Different Evaluation Methods In Jinjiang Town

As a important part of urban and rural construction land of our country,rural collective construction land is important survival carrier,which relating to farmers life closely.So,How to use rural collective construction reasonably not only affects the sustainable development of the country's social economy,but also affects the stability and prosperity of farmers,agriculture,rural,undering the current situation of protecting farmland strictly and insufficient urban construction land.In February 2015,Jinjiang town on Jiangxi province was taked the pilot of rural residential land reform,then it was taked the pilot ofentering market for the collectively operating land and land levying system reform too.This paper take Jinjiang Town on Jiangxi pilot of ruralreform county as the research object,grading rural collective construction land,basing on two different evaluation methods to explore the scientific and reasonable method that best suits the actual situation of Jinjiang town.The evaluation results are of great significance to guide the local rural land use,helpful for enriching the theoretical system of collective construction land gradation and standarding administration of rural collective construction land transformation.This paper evaluatesrural collective construction land grading based on the method of multi-factor comprehensive evaluation and principal component analysis.Firstly,rural collective construction land is divided four land grades,combing the method of multi-factor comprehensive evaluation and principal component analysis.Then,This paper divides the research area into a grid of cells 30m×30m using grid methods,choosing alternative grading factors from 7 aspects like business prosperity and traffic condition according experts demonstration.Then,Principal component analysis is used to determine the last 18 land grading factors and factors weight.According different attenuation mode to calculate index influence scores of each types.At the Last,the method of weighting summation is used to calculate the total grade of the unit,which between 29.36-89.51.The results showed that:the quantity of grades area was a trend of normal distribution,most of the units belonged to the second grade,which was 40.85%;the third grade taked the second place,accounting for 28.19%;the first grade had the smallest area,accounting for 13.20%;the fourth grade accounted for 17.76%.In spatial distribution,the center land of the town was the highest quality,and the grade of the land was gradually reduced from the center to the surroundings.Using BP neural network model for rural collective construction land grading again with the same evaluation system.This paper selects 2520 exercitations samples,631 test samples from four representative administrative villages evaluation grids and 11426 unknown samples from additional villages.Then,this paper built BP neural network model by setting input layer,output layer,the best number of hidden layer,learning rate,maximum training times and expectation error,achieving he target after 7 exercitations,which is6.6141e-07.At last,the exercised neural network model is employed to evaluate the sample of grade unknown to get the last reasults.The results showed that:in quantity,least of the units belonged to the first grade,which was 17.87%;the second grade taked the most place,accounting for 37.54%;the third and fourth grade had the approximate area,accounting for 23.93% and 20.66%.In spatial distribution,the center land of the town was the highest quality,and the grade of the land was reduced from the center to the surroundings,butthere was a phenomenon of skipping grades between neighboring villages.At the last,comparing the evaluation results for two different methods from quantity and spatial distribution,the result which combing the method of multi-factor comprehensive evaluation and principal component analysis was closer to the true value distribution than the BP Neural network model.So,The index system can reflect the unique regional characteristics of rural collective construction land truly,which basing on the difference between collective construction land and urban construction land.The land grading method improved the accuracy and objectivity of the rural collective construction land grading results and establishing basis for making technical norm which combing the method of multi-factor comprehensive evaluation and principal component analysis.In the meanwhile,This paper suggests that the method of multi-factor comprehensive evaluation guarantees the accuracy of the evaluation combing GIS space analysis,the method of BP Neural Network applies sufficient samples and fast evaluation for small area.