修改: PriceSystem/sell_price/zhuge0118A.py

This commit is contained in:
罗干
2018-06-06 10:56:58 +08:00
parent 1e012e0759
commit 0c8080cd3d
+54 -18
View File
@@ -745,7 +745,9 @@ def commonprice(ct, rl, p,weekend):
return cc1 return cc1
else: else:
return cc1 return cc1
#在同一个小区里,可能同时有别墅和普通住宅两种房源
#本函数就是根据本小区的物业类型提取房源
@timer @timer
def consider_villa(ct, rl, p): def consider_villa(ct, rl, p):
@@ -776,7 +778,7 @@ def commonprice(ct, rl, p,weekend):
return df return df
else: else:
return df return df
#本函数用于判断本小区是否是别墅,并提取别墅在DataFrame 中的index
def fetch_villa_index(dd): def fetch_villa_index(dd):
# dd=dd.dropna() # dd=dd.dropna()
@@ -801,6 +803,7 @@ def commonprice(ct, rl, p,weekend):
return villa return villa
#本函数就是提取进6个月以来某个小区第三方均价最新的一条
def dsfjujia(ct, rl, lk): def dsfjujia(ct, rl, lk):
query = {"city": ct, "region": rl, query = {"city": ct, "region": rl,
@@ -855,7 +858,7 @@ def commonprice(ct, rl, p,weekend):
aa_zz=set(junjia.index) aa_zz=set(junjia.index)
jiaji_lz= list(ll_zz & aa_zz) jiaji_lz= list(ll_zz & aa_zz)
""" """
#聚类算法
def k_means_clus(pp1, clus): def k_means_clus(pp1, clus):
from sklearn.cluster import KMeans from sklearn.cluster import KMeans
@@ -879,6 +882,11 @@ def commonprice(ct, rl, p,weekend):
else: else:
return int(pp1.mean()) return int(pp1.mean())
#判断两个群体是否可以协商出一个可行的价格
#junjia_lzf群体1
#junjia_anst,群体2
# ctr控制参数
#per控制百分比
def group(junjia_lzf, junjia_anst, ctr,per): def group(junjia_lzf, junjia_anst, ctr,per):
bb = [] bb = []
@@ -904,7 +912,7 @@ def commonprice(ct, rl, p,weekend):
bb.append([someprice, numb]) bb.append([someprice, numb])
###print('bb=',bb) ###print('bb=',bb)
return bb return bb
#两个群体是否可以协调出最优价格
def middle_operate(junjia_lzf, junjia_anst): def middle_operate(junjia_lzf, junjia_anst):
# #####print('%%%',ppd3['total_price']) # #####print('%%%',ppd3['total_price'])
@@ -962,7 +970,7 @@ def commonprice(ct, rl, p,weekend):
return aa1_6.iloc[0] return aa1_6.iloc[0]
else: else:
return None return None
#把第三方均价分成两个群体
def devide_group(al, bl, junjia): def devide_group(al, bl, junjia):
alls = set(junjia.index) alls = set(junjia.index)
@@ -977,7 +985,9 @@ def commonprice(ct, rl, p,weekend):
anst = set_div(bl, alls) anst = set_div(bl, alls)
return lzf, anst return lzf, anst
#本函数通过输入第三方均价和周边价格来计算小区均价
#junjia 是一个Series index 是网站的名称
#neighb 是周边小区的价格
def dglg_out1(junjia, neighb): def dglg_out1(junjia, neighb):
""" """
@@ -987,13 +997,14 @@ def commonprice(ct, rl, p,weekend):
'yungujia_esf_price','fanggugu_esf_price' 'yungujia_esf_price','fanggugu_esf_price'
] ]
""" """
#优质的样本源
al = [ al = [
'lianjia_esf_price', 'lianjia_esf_price',
'zhugezhaofang_esf_price', 'zhugezhaofang_esf_price',
'fanggugu_esf_price', 'fanggugu_esf_price',
'yungujia_esf_price' 'yungujia_esf_price'
] ]
#一般的样本源
bl = ['anjuke_esf_price', bl = ['anjuke_esf_price',
'soufang_esf_price', 'soufang_esf_price',
'tuboshi_esf_price', 'tuboshi_esf_price',
@@ -1002,6 +1013,7 @@ def commonprice(ct, rl, p,weekend):
] ]
mmax=junjia.max() mmax=junjia.max()
#次函数已经放弃采用
def junjia_max(mmax,dg12): def junjia_max(mmax,dg12):
if dg12!=None and mmax!=None: if dg12!=None and mmax!=None:
@@ -1020,7 +1032,7 @@ def commonprice(ct, rl, p,weekend):
# ppd3dp=ppd3['total_price'].dropna() # ppd3dp=ppd3['total_price'].dropna()
if type(neighb) != str and neighb != None: if type(neighb) != str and neighb != None:
#参考周边对junjia进行预处理
def cut_junjia(junjia,neighb): def cut_junjia(junjia,neighb):
if len(junjia)>3: if len(junjia)>3:
@@ -1034,6 +1046,7 @@ def commonprice(ct, rl, p,weekend):
junjia_lzf=junjia[al].dropna() junjia_lzf=junjia[al].dropna()
junjia_anst=junjia[bl].dropna() junjia_anst=junjia[bl].dropna()
''' '''
#对junjia 分成两个group
alldi = devide_group(al, bl, junjia) alldi = devide_group(al, bl, junjia)
junjia_lzf = alldi[0] junjia_lzf = alldi[0]
@@ -1133,6 +1146,8 @@ def commonprice(ct, rl, p,weekend):
#赶集和安居客几乎相同并且和其他数据不同,就drop赶集网 的均价 #赶集和安居客几乎相同并且和其他数据不同,就drop赶集网 的均价
judp=drop_ganji(junjia) judp=drop_ganji(junjia)
#通过第三方均价计算小区均价
dglg = dglg_out1(judp, neighb[0]) dglg = dglg_out1(judp, neighb[0])
print('dglg=',dglg) print('dglg=',dglg)
# dggg=dglg_out1(junjia,neighb[0]) # dggg=dglg_out1(junjia,neighb[0])
@@ -1141,6 +1156,7 @@ def commonprice(ct, rl, p,weekend):
# #####print('dggg=',dggg) # #####print('dggg=',dggg)
#tt=time.time() #tt=time.time()
#通过物业类型对房源进行分类
data = consider_villa(ct, rl, p) data = consider_villa(ct, rl, p)
#data.to_excel('data.xls') #data.to_excel('data.xls')
@@ -1151,10 +1167,11 @@ def commonprice(ct, rl, p,weekend):
####################################################### #######################################################
##print('rrrr') ##print('rrrr')
##print(data.shape[0]) ##print(data.shape[0])
#增加优质样本的权重
data=add_weight(data) data=add_weight(data)
##print('data=',data) ##print('data=',data)
if data.shape[0] > 3: if data.shape[0] > 3:
paa = add_block(data) paa = add_block(data)
ppd = paa[['total_price', 'area', 'region', 'district_name', 'url', 'block', 'p_date']] ppd = paa[['total_price', 'area', 'region', 'district_name', 'url', 'block', 'p_date']]
# #####print('ppd##############') # #####print('ppd##############')
@@ -1170,7 +1187,9 @@ def commonprice(ct, rl, p,weekend):
#ppd1[['total_price','area']].to_excel('ppd1.xls') #ppd1[['total_price','area']].to_excel('ppd1.xls')
ppd1['total_price'] = ppd1['total_price'] / ppd1['area'] ppd1['total_price'] = ppd1['total_price'] / ppd1['area']
#提取某个小区的区域价格
#一般情况下,用周边的价格对房源进行预处理,过滤那些极小值
#但有时候周边价格计算不出来,就需要用区域价格进行过滤
def take_region_price(ct,rl): def take_region_price(ct,rl):
@@ -1189,7 +1208,8 @@ def commonprice(ct, rl, p,weekend):
return -1 return -1
if ppd1.shape[0] > 0: if ppd1.shape[0] > 0:
#通过周边价格或区域价格对房源进行过滤
def guolv_data(ppd2, neighb): def guolv_data(ppd2, neighb):
if type(neighb) is not str: if type(neighb) is not str:
ph = ppd2[ppd2['total_price'] > 0.7 * neighb] ph = ppd2[ppd2['total_price'] > 0.7 * neighb]
@@ -1267,13 +1287,19 @@ def commonprice(ct, rl, p,weekend):
return -2, float('%.2f' % credit['0.85']), value_type['第三方'], sample_size return -2, float('%.2f' % credit['0.85']), value_type['第三方'], sample_size
else: else:
return -2,float('%.2f' % credit['0.85']), value_type['第三方'], sample_size return -2,float('%.2f' % credit['0.85']), value_type['第三方'], sample_size
#在计算均价时,同是参考第三方均价计算出的价格和房源价格
#dglg:第三方均价计算出的价格
def map_time_price(junjia,dglg,ppd3): def map_time_price(junjia,dglg,ppd3):
day_15=front_step_time(weekend,15) day_15=front_step_time(weekend,15)
day_30=front_step_time(weekend,30) day_30=front_step_time(weekend,30)
day_60=front_step_time(weekend,60) day_60=front_step_time(weekend,60)
#考虑最近15天的房价和dglg
price_15= consider_time_ss1(junjia,dglg,ppd3,day_15) price_15= consider_time_ss1(junjia,dglg,ppd3,day_15)
#考虑最近30天的房价和dglg
price_30= consider_time_ss1(junjia,dglg,ppd3,day_30) price_30= consider_time_ss1(junjia,dglg,ppd3,day_30)
#考虑最近60天的房价和dglg
price_60= consider_time_ss1(junjia,dglg,ppd3,day_60) price_60= consider_time_ss1(junjia,dglg,ppd3,day_60)
if price_15[0]>0: if price_15[0]>0:
@@ -1290,17 +1316,18 @@ def commonprice(ct, rl, p,weekend):
dggg=map_time_price(junjia,dglg,ppd3) dggg=map_time_price(junjia,dglg,ppd3)
#当第三方均价的样本量比较大,就可以得出一个非常准确的均价,故直接输出这个价格不考虑房源
#这一步计算出的价格可信度最高为0.88
if len(junjia)>4 and dglg!=None: if len(junjia)>4 and dglg!=None:
sample_size = len(junjia) sample_size = len(junjia)
return dglg, float('%.2f' % credit['0.88']), value_type['第三方'], sample_size return dglg, float('%.2f' % credit['0.88']), value_type['第三方'], sample_size
#这一步在第三方均价算出一个价格之后,并参考房源,计算均价,可信度0.85
elif dggg[0]>0 : elif dggg[0]>0 :
return dggg return dggg
#这个一步采用链家和诸葛找房房源聚类出一个价格
elif length > 3: elif length > 3:
# #####print('ldf=',l_and_z_and_i) # #####print('ldf=',l_and_z_and_i)
pppk = ppd3.loc[l_and_z, :] pppk = ppd3.loc[l_and_z, :]
@@ -1317,6 +1344,10 @@ def commonprice(ct, rl, p,weekend):
return int(gm), float('%.2f' % credit['0.8']), value_type['自身样本'], sample_size # 优质样本 return int(gm), float('%.2f' % credit['0.8']), value_type['自身样本'], sample_size # 优质样本
elif data.shape[0] > 0: elif data.shape[0] > 0:
data['avg'] = data['total_price'] / data['area'] data['avg'] = data['total_price'] / data['area']
@@ -1645,17 +1676,21 @@ def pt_up(ct, rl, p, weekend, c_date, cmd_estate_type):
}, True }, True
) )
#本函数用于将计算出的价格保存到数据库
@timer @timer
def out_put_price(ct, rl, lk, weekend): def out_put_price(ct, rl, lk, weekend):
#提取本小区的信息
lf = region_district_list_lk(ct, rl, lk) lf = region_district_list_lk(ct, rl, lk)
if lf.shape[0]>0: if lf.shape[0]>0:
lf = add_alias(lf) lf = add_alias(lf)
#p是一个列表['小区名','小区别名','物业类型']
p = ghhh(lf[['name', 'alias', 'estate_type2']])[0] p = ghhh(lf[['name', 'alias', 'estate_type2']])[0]
######print('p=', p) ######print('p=', p)
#t1=time.time() #t1=time.time()
#下面函数用于计算价格
mm = commonprice(ct, rl, p,weekend) mm = commonprice(ct, rl, p,weekend)
##print('mmjisuantime=',time.time()-t1,'yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy') ##print('mmjisuantime=',time.time()-t1,'yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy')
######print('mm=',mm) ######print('mm=',mm)
@@ -1685,7 +1720,8 @@ def out_put_price(ct, rl, lk, weekend):
#print('last_sunday_price=',last_sunday_price) #print('last_sunday_price=',last_sunday_price)
# #####print('mm[0]',mm[0]) # #####print('mm[0]',mm[0])
#计算小区的涨跌幅度
def make_up(avg,last_sunday_price): def make_up(avg,last_sunday_price):
if type(last_sunday_price) != str and last_sunday_price != None and mm[0] != None and type(mm[0]) != str: if type(last_sunday_price) != str and last_sunday_price != None and mm[0] != None and type(mm[0]) != str:
@@ -1698,7 +1734,7 @@ def out_put_price(ct, rl, lk, weekend):
######print('make_up',make_up,'*'*20) ######print('make_up',make_up,'*'*20)
######print('credic',mm[1]) ######print('credic',mm[1])
#对数据进行处理,控制其涨跌幅度
def avg_compare(ct,rl,lk,mm,weekend): def avg_compare(ct,rl,lk,mm,weekend):
avg=fix.price_fix.update(ct,rl,lk,int(mm[0]),weekend) avg=fix.price_fix.update(ct,rl,lk,int(mm[0]),weekend)
##print('nihe=',avg) ##print('nihe=',avg)
@@ -1714,7 +1750,7 @@ def out_put_price(ct, rl, lk, weekend):
#avg=int(mm[0]) #avg=int(mm[0])
#计算小区的涨跌幅度
make_up = int(make_up(avg,last_sunday_price)) make_up = int(make_up(avg,last_sunday_price))
##print('s1_time=',time.time()-s1) ##print('s1_time=',time.time()-s1)
#avg=mm[0] #avg=mm[0]