Research On The Application Of Data Mining Technology In Rental Data

With the substantial increase in house prices in recent years,people get more influenced by the restrictions on purchasing power,more and more people choose to rent to live,the size of China's housing rental market will become larger and larger,and the more business opportunities will be brought about.Due to the huge amount of rental information released daily,the rental data contains a lot of information,and only recently released rental data can be seen,it is impossible to directly observe the hidden laws behind these data and discover the potential value.This article is based on the data mining technology to analyze and study the rental housing data released by the shell housing search website for the whole year of 2019 in Kunming City,Chenggong District,to mine the laws and values in the rental housing data information,to provide tenants and businesses with realistic research analysis.In this paper,we analyze the rental housing data from two perspectives:A study of spatio-temporal clustering analysis of rental housing data is proposed.In this paper,the ST-DBSCAN clustering analysis model for rental data is constructed to address the data characteristics of rental data,including release time,geographical location information and attribute characteristics.First,a new frequency-based threshold setting method is proposed to address the shortcomings of the ST-DBSCAN algorithm,which relies on human experience to determine the threshold value.The data simulation experiment verifies that the newly proposed threshold setting method can improve the accuracy of clustering and can identify some low density data clusters to obtain more accurate clustering results.Secondly,a computational model of the property similarity of the rental data is constructed and applied to the ST-DBSCAN algorithm.The spatio-temporal clustering analysis of the rental data is divided into two aspects: first,the clustering analysis is conducted on the distribution laws in the temporal and spatial dimensions of the rental data released in the whole year of 2019,and then,by applying the attribute similarity computation model,the influence of the two attributes of "home appliance facilities" and "average price" on the spatio-temporal clustering results of the rental data is studied.An analytical study of correlations between rental housing data attributes is pro-posed.Spatio-temporal clustering analysis can only mine the distribution laws of rental housing data information in the temporal and spatial dimensions,and it cannot mine the value of the correlation relationship between data attributes and characteristics.In order to mine the association laws between various attribute features of rental housing information data,this paper applies the association rule mining algorithm to the analysis of rental housing data.The setting of two important parameters for the existence of the traditional association rule algorithm Apriori,support and confidence,relies on human determination and has a certain subjective influence on the generation of association rules.An improved CPS-Apriori association algorithm based on the chaotic particle swarm algorithm is proposed,which uses the chaotic particle swarm(CPS)to optimize the setting of support and confidence,shortening the rule extraction time and improving the operation efficiency of the algorithm.The CPS-Apriori association algorithm is applied to the association rule mining among rental data attributes to perform a research analysis of the association relationships among rental data attributes.