Skip to content

F
ffm-baseline
Commit d9ed33cc authored by 张彦钊's avatar 张彦钊
Browse files
Options

Merge branch 'master' of git.wanmeizhensuo.com:ML/ffm-baseline 

add filter cid
parents eef1f0c7 7be44615
Hide whitespace changes
Inline Side-by-side
Showing with 40 additions and 10 deletions
eda/recommended_indexs/main.py
View file @ d9ed33cc
......@@ -8,6 +8,13 @@ from func import *
print("开始获取特征数据...")
#1.0 question曝光占比(=question被曝光数/总cid被曝光数)
question_imp_rate_all = CidRate("all","question").get_cid_imp_rate("所有")
question_imp_rate_ios = CidRate("ios","question").get_cid_imp_rate("苹果")
question_imp_rate_android = CidRate("android","question").get_cid_imp_rate("安卓")
question_imp_rate_result = [question_imp_rate_all,question_imp_rate_ios,question_imp_rate_android]
print("已获取question曝光占比")
#1.1 answer曝光占比(=answer被曝光数/总cid被曝光数)
answer_imp_rate_all = CidRate("all","answer").get_cid_imp_rate("所有")
answer_imp_rate_ios = CidRate("ios","answer").get_cid_imp_rate("苹果")
......@@ -42,21 +49,28 @@ click_answer_android = ClkCidUidRate("android","answer").get_clk_cid_uid_rate("
click_answer_result = [click_answer_all,click_answer_ios,click_answer_android]
print("已获取点击answer用户占比")
#1.6 点击diary用户占比(=点击diary用户数/曝光diary用户数)
#1.6 点击question用户占比(=点击question用户数/曝光question用户数)
click_question_all = ClkCidUidRate("all","question").get_clk_cid_uid_rate("所有")
click_question_ios = ClkCidUidRate("ios","question").get_clk_cid_uid_rate("苹果")
click_question_android = ClkCidUidRate("android","question").get_clk_cid_uid_rate("安卓")
click_question_result = [click_question_all,click_question_ios,click_question_android]
print("已获取点击question用户占比")
#1.7 点击diary用户占比(=点击diary用户数/曝光diary用户数)
click_diary_all = ClkCidUidRate("all","diary").get_clk_cid_uid_rate("所有")
click_diary_ios = ClkCidUidRate("ios","diary").get_clk_cid_uid_rate("苹果")
click_diary_android = ClkCidUidRate("android","diary").get_clk_cid_uid_rate("安卓")
click_diary_result = [click_diary_all,click_diary_ios,click_diary_android]
print("已获取点击diary用户占比")
#1.7 有点击用户占比(=有点击用户数/有曝光用户数)
#1.8 有点击用户占比(=有点击用户数/有曝光用户数)
click_everything_all = ClkCidUidRate("all","everything").get_clk_cid_uid_rate("所有")
click_everything_ios = ClkCidUidRate("ios","everything").get_clk_cid_uid_rate("苹果")
click_everything_android = ClkCidUidRate("android","everything").get_clk_cid_uid_rate("安卓")
click_everything_result = [click_everything_all,click_everything_ios,click_everything_android]
print("已获取有点击用户占比")
#1.8 无点击用户数分布(=无点击用户∩激活用户 / 激活用户数) #注意：(]里面的数字指的是距离当前时间的天数
#1.9 无点击用户数分布(=无点击用户∩激活用户 / 激活用户数) #注意：(]里面的数字指的是距离当前时间的天数
try:
click_zero_uid_detail_all = get_click_zero_uid_rate_detail("all")
click_zero_uid_detail_all["platform"] = "所有"
......@@ -107,14 +121,16 @@ def result2file():
line = """数据日期：{}
内容概览：以下所有数据都是昨天一天的首页的
1. 比例特征
1.0 question曝光占比(=question被曝光数/总cid被曝光数)
1.1 answer曝光占比(=answer被曝光数/总cid被曝光数)
1.2 diary曝光占比(=diary被曝光数/总cid被曝光数)
1.3 活跃用户点击率(=有点击用户点击次数/有点击用户曝光次数)
1.4 活跃用户平均每天曝光次数(=活跃用户曝光次数/独立活跃用户数)
1.5 点击answer用户占比(=点击answer用户数/曝光answer用户数)
1.6 点击diary用户占比(=点击diary用户数/曝光diary用户数)
1.7 有点击用户占比(=有点击用户数/有曝光用户数)
1.8 无点击用户数分布(=无点击用户∩激活用户 / 激活用户数) #注意：平台一排的数字指的是距离当前时间的天数
1.6 点击question用户占比(=点击question用户数/曝光question用户数)
1.7 点击diary用户占比(=点击diary用户数/曝光diary用户数)
1.8 有点击用户占比(=有点击用户数/有曝光用户数)
1.9 无点击用户数分布(=无点击用户∩激活用户 / 激活用户数) #注意：平台一排的数字指的是距离当前时间的天数
2.Top特征
2.1 用户点击次数分布(第一列：用户点击次数；第二列：独立用户数量)
2.2 Top 100 diary (sorted by ctr)
......@@ -128,6 +144,12 @@ def result2file():
f.write(line)
f.write("#1. 比例特征\n")
f.write("=================================================================\n")
f.write("#1.0question曝光占比(=question被曝光数/总cid被曝光数)\n")
f.write(tplt.format("平台","question被曝光数","总cid被曝光数","question被曝光数占比"))
for i in question_imp_rate_result:
line = tplt.format(i[0],i[1],i[2],"{}%".format(round(i[3]*100,2)))
f.write(line)
f.write('\n')
f.write("#1.1answer曝光占比(=answer被曝光数/总cid被曝光数)\n")
f.write(tplt.format("平台","answer被曝光数","总cid被曝光数","answer被曝光数占比"))
for i in answer_imp_rate_result:
......@@ -158,20 +180,26 @@ def result2file():
line = tplt.format(i[0],i[1],i[2],"{}%".format(round(i[3]*100,2)))
f.write(line)
f.write('\n')
f.write("#1.6点击diary用户占比(=点击diary用户数/曝光diary用户数)\n")
f.write("#1.6点击question用户占比(=点击question用户数/曝光question用户数)\n")
f.write(tplt.format("平台","点击question用户数","曝光question用户数","击question用户占比"))
for i in click_question_result:
line = tplt.format(i[0],i[1],i[2],"{}%".format(round(i[3]*100,2)))
f.write(line)
f.write('\n')
f.write("#1.7点击diary用户占比(=点击diary用户数/曝光diary用户数)\n")
f.write(tplt.format("平台","点击diary用户数","曝光diary用户数","击diary用户占比"))
for i in click_diary_result:
line = tplt.format(i[0],i[1],i[2],"{}%".format(round(i[3]*100,2)))
f.write(line)
f.write('\n')
f.write("#1.7有点击用户占比(=有点击用户数/有曝光用户数)\n")
f.write("#1.8有点击用户占比(=有点击用户数/有曝光用户数)\n")
f.write(tplt.format("平台","have点击用户数","have曝光用户数","have点击用户占比"))
for i in click_everything_result:
line = tplt.format(i[0],i[1],i[2],"{}%".format(round(i[3]*100,2)))
f.write(line)
f.write('\n')
if "click_zero_uid_detail_result" in dir():
f.write("#1.8无点击用户数分布(=无点击用户∩激活用户 / 激活用户数) #注意：(]里面的数字指的是距离当前时间的天数\n")
f.write("#1.9无点击用户数分布(=无点击用户∩激活用户 / 激活用户数) #注意：(]里面的数字指的是距离当前时间的天数\n")
f.write("平台"+'\t\t'+"0-7"+'\t\t'+"7-14"+'\t\t'+ \
"14-30"+'\t\t'+"30-60"+'\t\t'+"60-90"+'\t\t'+"90+"+'\n')
for i in click_zero_uid_detail_result:
......@@ -243,7 +271,9 @@ def rate2file():
str(activate_uid_imp_all[3])+','+str(activate_uid_imp_beijing[3])+','+\
str(click_answer_all[3])+','+str(click_answer_ios[3])+','+str(click_answer_android[3])+','+\
str(click_diary_all[3])+','+str(click_diary_ios[3])+','+str(click_diary_android[3])+','+\
str(click_everything_all[3])+','+str(click_everything_ios[3])+','+str(click_everything_android[3])+'\n'
str(click_everything_all[3])+','+str(click_everything_ios[3])+','+str(click_everything_android[3])+','+\
str(question_imp_rate_all[3])+','+str(question_imp_rate_ios[3])+','+str(question_imp_rate_android[3])+','+\
str(click_question_all[3])+','+str(click_question_ios[3])+','+str(click_question_android[3])+'\n'
f.write(line)
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment