import os import random import sys from pyspark import SparkConf, SparkContext from pyspark.sql import SQLContext APP_NAME = "prediction" stock_cnt = 9 # Load and parse the data def parsePoint(ss): line = ss.split(" ") values = [float(x) for x in line] return LabeledPoint(values[0], values[1:]) if __name__ == "__main__": conf = SparkConf().setAppName(APP_NAME) conf = conf.setMaster("local[*]") sc = SparkContext(conf=conf) model_param = sc.textFile("model_param") pred_data = sc.textFile("pred_input") # Build the model model = LinearRegressionWithSGD.setModel(model_param) # Evaluate the model on training data pred_res = pred_data.map(lambda fea: model.predict(fea)) pred_res.saveAsTextFile("pred_res")