Day 14_house_price

# -*- coding: utf-8 -*-
"""Day 20_house_price.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/13AQ4xoIUMEpJA1zQHEV95Ebhh48pHccm
"""
 
from google.colab import drive
drive.mount('/gdrive')
 
PATH = "/gdrive/My Drive/Colab Notebooks/resources/"
 
import matplotlib.pyplot as plt
from tensorflow import keras
import tensorflow as tf
import numpy as np
import pandas as pd
 
print(tf.__version__)
 
df = pd.read_csv(PATH + "house/train.csv")
 
df[:2]
 
df = df[['sqft_living','bedrooms', 'bathrooms', 'sqft_lot','floors','zipcode','price']]
 
np.sum(df.isna())
 
## 카테고리로 바꿀만한 컬럼 확인
df.apply(lambda x : x.value_counts().size, axis = 0)
 
df[['bedrooms',"bathrooms", 'floors', 'zipcode']] = df[['bedrooms',"bathrooms", 'floors', 'zipcode']].astype('category')
 
df2 = pd.get_dummies(df,prefix = ['bedrooms',"bathrooms", 'floors', 'zipcode'])
 
df2[:2]
 
## 나중에 복원용
target_median = np.percentile(df2['price'], 50)
target_Q1     = np.percentile(df2['price'], 25)
target_Q3     = np.percentile(df2['price'], 75)
target_IQR    = target_Q3 - target_Q1
 
 
df.apply(lambda x : x.value_counts().size, axis = 0)
 
from sklearn.preprocessing import RobustScaler
df3 = df2.copy()
df3[['sqft_living','sqft_lot','price']] = RobustScaler().fit_transform(df2[['sqft_living','sqft_lot','price']])
 
df3[:2]
 
ydata  = df3['price']
xdata  = df3.drop(['price'],axis = 1)
 
from sklearn.model_selection import train_test_split
xtrain, xtest, ytrain, ytest = train_test_split(xdata, ydata)
 
xtrain.shape , xtest.shape, ytrain.shape, ytest.shape
 
 
 
ytrain = np.array(ytrain).reshape(-1,1)
ytest  = np.array(ytest).reshape(-1,1)
 
x = tf.placeholder(tf.float32, shape = (None, 118))
y = tf.placeholder(tf.float32, shape = (None, 1))
 
w = tf.Variable(tf.random.normal([118,1]), name = "weight")
b = tf.Variable(tf.random.normal([1])  , name = 'bias')
 
hf = x @ w + b
cost = tf.reduce_mean(tf.square(hf - y)) 
 
train = tf.train.GradientDescentOptimizer(1e-3).minimize(cost)
 
sess = tf.Session()
sess.run(tf.global_variables_initializer())
 
for step in range(50000):
    cv, hv, _ = sess.run([cost, hf, train], feed_dict = {x : xtrain , y : ytrain})
    
    if step % 500 == 0:
        print(step, "cost : ", cv, "prediction : ", hv)
        if cv < 0.001:
            break
 
import matplotlib.pyplot as plt
import seaborn as sns
 
plt.scatter(x = hv, y = ytrain)
xy = np.linspace(-30, 10, 41)
sns.lineplot(x = xy, y = xy)
plt.xlim(-5, 10)
 
predict = sess.run([hf], feed_dict = {x : xtest })
 
## 예측
predict2 = predict[0] * target_IQR + target_median
ytest2   = ytest * target_IQR + target_median
 
plt.scatter(x = predict2, y = ytest2)
xy = np.linspace(-0, 35e+10, 2)
sns.lineplot(x = xy, y = xy)
 
predict2.shape
 
ytest.shape