论文信息 - Construction of Training Data for Price Prediction of a Real Estate from Internet Ads

Construction of Training Data for Price Prediction of a Real Estate from Internet Ads

The paper presents a model for constructing a data set aimed at predicting a price of a real estate (houses and flats) from the standard Internet ads. The model for predicting a real estate price includes, in addition to standard real estate's features (area, number of bedrooms, etc.) appearing in ad, attractiveness of a real estate location as well as information on some additional interior facilities (e.g., refrigerator, dish-washing machine, stove, etc.). The proposed training set construction model uses OpenStreetMap's Overpass API for determining attractiveness of a real estate's location, and a convolution neural network for detecting interior facilities from real estate photos.

[1] J. Friedman. Greedy function approximation: A gradient boosting machine. , 2001 .

[2] John R. Ottensmann,et al. Urban Location and Housing Prices within a Hedonic Model , 2008 .

[3] N. Nakagoshi,et al. Using GIS and landscape metrics in the hedonic price modeling of the amenity value of urban green space: A case study in Jinan City, China , 2007 .

[4] Pietro Perona,et al. Microsoft COCO: Common Objects in Context , 2014, ECCV.

[5] Yoav Freund,et al. A decision-theoretic generalization of on-line learning and an application to boosting , 1995, EuroCOLT.

[6] Sergio Guadarrama,et al. Speed/Accuracy Trade-Offs for Modern Convolutional Object Detectors , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[7] Jian Sun,et al. Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).