﻿ 基于集成学习的房租预测研究

# 基于集成学习的房租预测研究Research of Prediction on House Rent Based on Intergration Learning

Abstract: The rapid development of the housing rental market has led to an increasing demand for housing rental information. There is always a problem of information asymmetry at both ends of the rental market. The rent is determined by many factors together. Accuracy of a single prediction model is unstable and is often affected by factors such as model performance, noise, and over-fitting risk. This study aims to develop and evaluate models of rental market dynamics using stacking integra-tion strategy on data from the DC competition community. We use the three basic models of Ran-dom Force Regressor, Extra Trees Regressor and LightGBM and establish a rent prediction model for integrated learning. The experimental results show that the prediction accuracy of this method is obviously better than any single prediction model, which improves the accuracy and stability of the prediction, and confirms the validity of the model in rent prediction.

1. 引言

2. 相关研究

3. 研究思路和方法

3.1. 问题分析

Figure 1. Solution of the problem

3.2. 模型比较

Table 1. Model comparison

3.3. 集成学习策略

Stacking是一种非线性的融合决策,是一种从原数据集中自动抽取有效特征的表示学习。一般来说Stacking就是训练一个多层的学习器结构，第一层称为学习层，用n个不同的分类器，将得到的预测结果合并为新的特征集，并作为下一层分类器的输入，通过第二层的输出训练器得到最终预测结果。为了防止过度拟合问题，Stacking在第一层模型训练时采用K折交叉检验的方式，第二层输出训练器一般选用逻辑回归模型。

Figure 2. Stacking strategy

4. 数据

Table 2. Feature description

4.1. 数据预处理

4.2. 特征工程

Figure 3. Feature correlation analysis

Table 3. Characteristic construction

5. 模型的建立

5.1. 单一模型训练

Table 4. RF Optimal parameters

Table 5. ET Optimal parameters

Table 6. Lgb Optimal parameters

Table 7. Scores of each model

5.2. 集成学习

Table 8. Mean square error and mean absolute error of each model

6. 总结与展望

[1] 中国软件行业协会培训中心. 2018年全国大学生计算机技能应用大赛[EB/OL]. http://www.cnccac.com/, 2018-8-20.

[2] 郑文娟. 中国城市住房价格与住房租金的影响因素及相互关系研究[D]: [博士学位论文]. 浙江: 浙江大学, 2011.

[3] 陈思翀, 陈英楠. 中国住房市场波动的影响因素研究——基于租金收益率的方差分解[J]. 金融研究, 2019, 464(2): 140-157.

[4] Li, J.Z. (2018) Monthly Housing Rent Forecast Based on LightGBM (Light Gradient Boosting) Model. In-ternational Journal of Intelligent Information and Management Science, 7, 6.

[5] Ma, Y., Zhang, Z., Ihler, A. and Pan, B. (2018) Estimating Warehouse Rental Price Using Machine Learning Techniques. International Journal of Computers Com-munications & Control, 13, 235-250
https://doi.org/10.15837/ijccc.2018.2.3034

[6] Wang, J.J., Hu, S.G., Zhan, X.T., et al. (2018) Predicting House Price with a Memristor-Based Artificial Neural Network. IEEE Access, 6, 6.
https://doi.org/10.1109/ACCESS.2018.2814065

[7] Mu, J., Wu, F. and Zhang, A. (2014) Housing Value Forecasting Based on Machine Learning Methods. Abstract and Applied Analysis, 2014, Article ID: 648047.
https://doi.org/10.1155/2014/648047

[8] 李春生, 李霄野, 张可佳. 基于遗传算法改进的BP神经网络房价预测分析[J]. 计算机技术与发展, 2018, 28(8): 144-147.

[9] Noor, K. and Jan, S. (2017) Vehicle Price Prediction System Using Machine Learning Techniques. International Journal of Computer Applications, 167, 27-31.
https://doi.org/10.5120/ijca2017914373

Top