Improved Prediction Accuracy of House Price Using Decision Tree Algorithm over Linear Regression Algorithm

Abstract

As a direct outcome of this research, it is planned that the accuracy of house price projections will be enhanced by using a novel decision tree algorithm rather than linear regression. This will be done in order to achieve the desired result (LR). The N=10 iteration of the Decision Tree Algorithm is put to use in order to generate the prediction. The size of the sample is figured out with the use of a G power Calculator, and a cutoff of 80% is decided upon as the minimum need for sufficient analytical power. The Linear Regression Method can be found in Group 1, whereas the New Decision Tree Algorithm can be found in Group 2. The confidence interval for the pre-test power is from 95% to 80%, the alpha value is 0.05, the beta value is 0.2, and the total number of participants in the study is twenty. In contrast, the accuracy of the New Decision Tree (DT) Algorithm was 90%, while the accuracy of the Linear Regression Algorithm was 80%. The findings of the statistical analysis that was carried out with the assistance of SPSS showed that the value of accuracy was insignificant: p=0.618 (p>0.05). The Innovative Decision Tree Algorithm outperforms the Linear Regression approach when it comes to estimating the value of real estate in the future.