| As the rapid development of big data in urban life,Big Data has provided us convenience in life.Meanwhile the application of urban planning and smart cities is increasing year by year.Urban mathematics such as computer science and urban planning has become a research hotspot in emerging fields.With the creation of open data,the use of city calculations to predict and plan urban development routes can accelerate the development of urban intelligence and improve the quality of life of residents.With the deterioration of the ecological environment,the concept of a livable city has gradually entered people’s lives.The livable city is a comprehensive evaluation of the suitable living conditions of the city.It is necessary to comprehensively consider the urban environment,life,and security,among which the convenience of living is appropriate.Habitat city evaluation in the important factors.The current assessment of the convenience of city life is mainly the use of questionnaire survey methods,which require a large amount of labor and take a long time.In this paper,by using the random forest algorithm based on open data to evaluate the convenience of city life,we have realized the algorithm of urban life convenience evaluation based on machine learning.In this paper,which analyzes the factors affecting the convenience of life,constructs an evaluation index system,and takes the statistical situation of the pedestrian accessibility of the surrounding facilities in all districts of the urban area to study the convenience of urban life.This essay is composed of several chapters.First,in data acquisition,crawler technology was used to obtain Baidu map poi data of urban open data.The Baidu map POI covers the latitude and longitude coordinates,names,and addresses of various facilities in the city.Secondly,in terms of feature extraction,the POI data of residential areas are filtered and filtered to be used as the cell data set,and the urban residential area data is used as the starting point.For the surrounding facilities of the community,the data set of the surrounding area of the community is obtained by rough estimation of the distance to the earth.Meanwhile the attributes of the urban area are extracted.Finally,Random Forest is better than the traditional algorithm.Random forest is better for fitting noise and small size data.The algorithm uses grid search and crossvalidation to optimize the model parameters.The algorithm is selected to be the bestfitting performance model to build a convenience assessment model.What is analyzed the importance of the evaluation feature;It is applied to the prediction administrative region through evaluation algorithm migration.For convenience of living,compare and analyze the convenience results of administrative districts.After research,it has been found that the application of random forests to predict the accuracy of the data is higher than other algorithms in accuracy and F1 value.The accuracy rate reaches 86%,which proves that the random forest is better for the noise data processing due to the introduction of two randomness.In the application of evaluation algorithms to promote the convenience of living in the administrative area,the results show that there is a high correlation between the convenience of living in the administrative area and the average price of the housing in the administrative area.The location factor is the main factor affecting the housing price,and the evaluation algorithm is reflected in high correlation,or more can from a side,reflected the algorithm predictive ability.Using open data and random forest model as research methods.It can quickly and easily extract evaluation indicators and establish efficient evaluation algorithms.It also explores a new way of thinking and new methods for life convenience analysis,which fills in the convenience decision-making evaluation of urban life based on open data.The blank of support provides a practical reference for the planning of urban living service facilities. |
PDF Full Text Request