Participation to the Yelp Competition of Kaggle

DATASET

Training

Restaurants: 2000 Images: 234843

Test

Restaurants: 10000 Images: ~237000

DEPENDENCIES

Dependencies for SURF & Clustering (http://mulan.sourceforge.net/)

• python 3.x
• openCV 2.4
• [scikit-learn] (http://scikit-learn.org/stable/)

(to make python 3 and openCV work for SURF follow the awesome instructions in this [link] (http://www.pyimagesearch.com/2015/07/16/where-did-sift-and-surf-go-in-opencv-3/))

Dependencies for Classification

• [weka 3.7.10] (https://sourceforge.net/projects/weka/files/weka-3-7/3.7.10/)
• java 1.7 or higher
• [mulan 1.5] (http://mulan.sourceforge.net/)

PROCEDURE:

• Load images
• Find POI (SIFT / SURF & OpenCV) & form POIVector
• Clustering on POI in order to form PhotoVector
• Clustering on Images in order to get form BusinessVector
• Form training dataset from BusinessVector and labels
• Classification & evaluation on training dataset
• Predictions on test dataset