sklearn房源数量预测demo

[button color="primary" icon="" url="https://own.masheng.fun/data/nanjing.csv" type=""]nanjing.csv[/button]

import numpy as np
import pandas as pd

names=['city','addr','rule','price','size','additional','zdate','unit']
df=pd.read_csv('/data/nanjing.csv',names=names)
df.head()

	city	addr	rule	price	size	additional	zdate	unit
0	南京	鼓楼-青云巷	3室1厅1卫	900	12㎡	电视冰箱洗衣机空调热水器床暖气宽带衣柜天然气	2020-03-08	元/月(季付价)
1	南京	鼓楼-南东瓜市	2室0厅1卫	1600	20㎡	NaN	2020-03-05	元/月
2	南京	鼓楼-五塘和园	4室1厅1卫	1390	12㎡	冰箱洗衣机空调热水器床宽带衣柜	2020-03-05	元/月(月付价)
3	南京	鼓楼-凤凰二村	1室1厅1卫	2000	27㎡	NaN	2020-03-06	元/月
4	南京	鼓楼-中海桃源里	4室1厅1卫	1090	10㎡	冰箱洗衣机空调热水器床宽带衣柜	2020-03-08	元/月(月付价)

df = df.fillna(method='ffill')

df[df.isnull().T.any()]

	city	addr	rule	price	size	additional	zdate	unit

hall=df['rule'].str.extract('.(\d)')
room=df['rule'].str.extract('(\d)')
room=room.astype('int64')
hall=hall.astype('int64')

df['rule']=(room+(hall*2))/2

df['size']=df['size'].str.extract('(\d+)')
df['size']=df['size'].astype('int64')

def additionals(a):
    counts=[]
    counts=a.split(' ')
    return len(counts)
df['additional']=df['additional'].apply(additionals)

data=df.groupby(df['zdate']).mean()
data.tail()

	rule	price	size	additional
zdate
2020-03-05	2.408271	1937.629323	49.803008	7.490977
2020-03-06	2.412957	1918.367250	52.073132	7.460254
2020-03-07	2.332008	2004.573890	50.938370	7.460570
2020-03-08	2.511217	1770.254929	35.565942	7.208702
2020-03-09	2.722754	1530.810778	30.033533	7.217964

counts=df.groupby('zdate').size()
counts.tail()

zdate
2020-03-05    1330
2020-03-06    1258
2020-03-07    1509
2020-03-08    2942
2020-03-09     835
dtype: int64

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

X=data
y=counts

X_train,X_test,y_train,y_test = train_test_split(X,y,test_size = 0.2,random_state=125)

## 建立线性回归模型
clf = LinearRegression().fit(X_train,y_train)
print('建立的LinearRegression模型为：','\n',clf)

建立的LinearRegression模型为： 
 LinearRegression(copy_X=True, fit_intercept=True, n_jobs=1, normalize=False)

y_pred = clf.predict(X_test)
print(len(y_pred))
print('预测前20个结果为：','\n',y_pred[:20])

26
预测前20个结果为： 
 [ 64.04629768  72.32580809 266.94981422 254.43703779 134.24151951
   0.80319776 122.11897184  70.2876959    5.53085201 232.45575409
 185.17606994 225.12269869  87.68120282 221.29098145 150.57651981
 -27.12364545 522.41034522 203.97223093 203.90648176  66.07014051]