Research On Housing Liquidity Based On Statistical Learning Methods

The ability that housing changes from the physical form to the value form(liquidity for short)is housing liquidity.In China's housing market,there has not been a systematic and accurate measurement index and analysis method of housing liquidity.In the traditional study,time on market what is the duration from listing date until time of sale about housing units is used as the measure of housing liquidity,and the housing search model is used as the academic model commonly used in housing liquidity research.In this paper,we use the transaction cycle(time on market)as the measure of housing liquidity,and turn it into discrete variable whose values are “good” and “bad” about housing liquidity as the output variable.Through the study of influence factors of housing liquidity,the factors influencing housing liquidity are selected as input variables.Then,according to the selected features,web crawler technology is used to get data for research and analysis from related websites.Finally,based on statistical learning methods,we use the obtained data to build a C5.0 classification prediction model.The research found there was a great relationship between housing liquidity and business circle,housing area.Factors like the floor where the housing unit is located,orientation of housing,there are schools or subway about housing or not,and so on are also very important to housing liquidity.The classification prediction model of housing liquidity has important value and significance to both test theory and solve practical problems.