机器学习 实验一 杭州二手房价预测

机器学习 实验一 杭州二手房价预测

一、实验环境

PC机，Python

二、代码

#%%

import pandas as pd
import random
import numpy as np
import matplotlib.pyplot as plt

#%%

train_data = pd.read_csv('train.csv')

#%%

train_data

#%%

test_data = pd.read_csv('test.csv')

#%%

test_data

#%%

full = train_data.append(test_data,ignore_index=True)

#%%

full.columns

#%%

full.head()

#%%

full.info()

#%%

full['单价'].unique()

#%%

full['楼型'].unique()

#%%

full.drop(['Unnamed: 0','简介','单价'], axis=1, inplace=True)
full

#%%

# room=[]
# living_room=[]
# for i in full['规格']:
#     if(i=='车位'):
#         room.append(0)
#         living_room.append(0)
#     else:
#         for j in range(len(i)):
#             if i[j]=='室':
#                  b=j
#         room.append(i[:b])
#         living_room.append(i[b+1:b+2])
# room
# len(room)

#%%

# len(living_room)

#%%

# renovation=[]
# for i in full['装修']:
#     if(i=='精装'):
#         renovation.append(0)
#     elif(i=='简装'):
#         renovation.append(1)
#     elif(i=='毛坯'):
#         renovation.append(2)
#     elif(i=='其他'):
#         renovation.append(3)
#     elif(i=="板塔结合"):
#         renovation.append(4)
#     elif(i=='板楼'):
#         renovation.append(5)
# renovation
# len(renovation)

#%%

# full['装修'].unique()

#%%

# building_type=[]
# for i in full['楼型']:
#     if(i=='板塔结合'):
#         building_type.append(0)
#     elif(i=='板楼'):
#         building_type.append(1)
#     elif(i=='塔楼'):
#         building_type.append(2)
#     elif(i=='暂无数据'):
#         building_type.append(3)
#     elif(i=='平房'):
#         building_type.append(4)
#     else:
#         building_type.append(np.random.randint(0,5))
# building_type
# len(building_type)

#%%

# full['室']=room
# full['厅']=living_room
# # full['装修']=renovation
# # full['楼型']=building_type
# full.drop(['规格'],axis=1, inplace=True)

#%%

for i in range(21318):
    full['总价'][i]=full['总价'][i][:-1]
full['总价']

#%%

for i in range(31318):
    full['大小'][i]=full['大小'][i][:-2]
full['大小']


#%%

full.info()


#%%

full['大小']=full['大小'].astype(float)
full['总价']=full['总价'].astype(float)
# full['室']=full['室'].astype(float)
# full['厅']=full['厅'].astype(float)

#%%

full.info()

#%%

full

#%%

object_type=[]
for col in full.columns:
    if full[col].dtype =="object":
        object_type.append(col)
full_dummied_object = full
for ob in object_type:
    objectDf=pd.DataFrame()
    objectDf=pd.get_dummies(full_dummied_object[ob],prefix=ob)
    full_dummied_object=pd.concat([full_dummied_object,objectDf],axis=1)
    full_dummied_object.drop(ob,axis=1,inplace=True)
    full_dummied_object.head()
full=full_dummied_object

#%%

full


#%%

from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn import linear_model

lr = linear_model.LinearRegression()

train_data=full[:21318]
source_y=train_data['总价']
train_copy=train_data.copy()
train_copy.drop(['总价'],axis=1,inplace=True)
source_x=train_copy

#%%

source_x


#%%

source_y

#%%

from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
maxscore=0
for i in range(100):
    train_X, test_X, train_y, test_y = train_test_split(source_x, source_y, train_size=0.8)
    model = LinearRegression()
    model.fit(train_X , train_y)
    LinearRegression(copy_X=True, fit_intercept=True, n_jobs=None, normalize=False)
    print(model.score(test_X, test_y))
    if(model.score(test_X, test_y)>maxscore):
        maxscore=model.score(test_X, test_y)
        modelbest=model
    if(maxscore>0.8):
        break
maxscore

#%%

test_data=full[21318:]
test_copy=test_data.copy()
test_copy.drop(['总价'], axis=1, inplace=True)
test_copy

#%%

result=modelbest.predict(test_copy)
result

二、实验结果与分析

1、猎豹平台提交结果：
2、一开始可以尝试对“地区”、“规格”、“大小”、“装修”、“楼型”、“总价”这几个特征进行处理，进行预测；为进一步提高正确率考虑逐步加入其他几个特征，并进行独热编码。