ayee / car-price-prediction-system Goto Github PK
View Code? Open in Web Editor NEWThis project forked from aniketvjadhav/car-price-prediction-system
Technologies: Nutch 1.6, MapReduce in Java, Mahout. For selling a used car, its price can be predicted by giving some attributes. e.g. Car Model, total miles, engine type. Trained data will be provided to our system to predict the price for new data. Used Nutch to crawl car data from 52 states which is on craigslist.org. Over 0.3 million records were fetched. The content was then pruned using two MapReduce Jobs. The first MapReduce cleaned the data removing unwanted unicode symbols and incomplete data (information without car model or total miles or engine or price). The Second MapReduce extracted the required attributes and emitted in tsv format. This tsv was then provided to a Naïve Based Classifier in Mahout. A classification model was built from the training data. This models predicts the price when attributes like car model model, miles, engine was provided.