The use of regression trees to the analysis of real estate market of housing

Abstract

Housing property can be described using a number of traits, some of which turn out to be difficult to express in numerical form. You can also use parameters which are not subject to explicit quantification. Belong to it selected location or membership of an exclusive building project, or opposed to a specific buyer reluctance of the complex. The attributes relevant to buyers may be disregarded just because of problems with their recognition in shaping market prices. A multitude of real estate tenancy causes difficulties with their comparison. If you make this scaling problem adopted features such analysis may not be sufficient. An attempt to include as many information may be restricted by the form of analysis. The undoubted advantage of the classification tree is the ability to delineate the characteristics of a quantitative (continuous or discrete) on a par with those of a qualitative nature (from the nominal scale after an ordinal). They allow an assessment of the impact of both the characteristics of qualitative, as well as a quantitative variable quantitative, without having to specify an arbitrary numeric values for the variables on the nature of the quality. This option allows you to include attributes such as subdivision, street name. Among the many methods of data analysis, it is worth noting the use of regression task for classification trees, so that the designation of property prices was possible based on the allocation to one of the nodes in the schema that you created earlier. The proposed method allows for the examination of the impact on the predictors of the dependent variable, as well as on the distribution of the existing set of homogeneous groups in terms of price.