import string
from math import log
import numpy as np


#create the dataset
#def createdataset():
#    lst=[map(float, line.strip().split()) for line in open("1.txt","r").readlines()]
#    lab=[line.strip().split() for line in open("2.txt","r").readlines()]
#    return lst,lab

#def createTree(dataSet,labels):

from numpy import * 
from numpy.linalg import *
# load txt
def training():
    data = loadtxt('1.txt',dtype=float)
    #data = np.array([map(float, line.strip().split(',')) for line in open("2.txt","r").readlines()])
    y = data[:,-1:]
    x = data[:,0:-1]
    x_t = x.transpose()
    
    tmp = matrix(x_t) * matrix(x)
    tmp1 = inv(tmp)
    result = matrix(tmp1) * matrix(x_t) * matrix(y) 
    return result

def testing():
    data = loadtxt('1test.txt',dtype=float)
    #data = np.array([map(float, line.strip().split(',')) for line in open("2test.txt","r").readlines()])
 
    y = data[:,-1:]
    x = data[:,0:-1]
    row, col = y.shape
    sum = 0
    fun = training()
    count = 0
    for i in range(row):
        delta_y = dot(x[i],fun) - y[i]
        delta_y = delta_y**2
        sum += delta_y
        count += 1
    return sum/count

print testing()