data_basic_stat

basic data exploration for JData

user data

• user data row number: 103616
• user_id distinct number:103616
• user age distribution:

index	user_age	user_cnt
0	0	12803
1	1	6
2	2	7999
3	3	46525
4	4	30828
5	5	3407
6	6	2048

- user sex distribution:

index	sex	user_cnt
0	0	45547
1	1	7585
2	2	50484

- user_lv_cd distribution:

index	user_lv_cd	user_cnt
0	1	2328
1	2	7519
2	3	21689
3	4	32205
4	5	39875

product data

• product data row number: 24187
• sku_id distinct number:24187
• sku_id distinct attr1:4

index	attr1	product_cnt
0	-1	1701
1	1	4760
2	2	3582
3	3	14144

- sku_id distinct attr2:3

index	attr2	product_cnt
0	-1	4050
1	1	13513
2	2	6624

- sku_id distinct attr3:3

index	attr3	product_cnt
0	-1	3815
1	1	8394
2	2	11978

- sku_id distinct cate:1

index	cate	product_cnt
0	8	24187

- sku_id distinct brand:102

comment data

• comment data row number: 558552
• sku_id distinct number:46546
• has_bad_comment distribution:

index	has_bad_comment	comment_cnt
0	0	292978
1	1	265574

action data

• action data row number: 79597328
• action data user_id number: 103565
• action data sku_id number: 31485
• action data sku_id in product number: 4284
• sku_id has more than one cate of brand: No
• action data cate=8 distinct sku_id number: 4284
• action data cate=8 sku_id not in product number: 0
• 20160401 last 5 day cate=8 buy distinct user_id number: 811
• 20160401 last 5 day cate=8 buy distinct

训练样本的数量

不做滑动窗口提取样本时

• offline train samples:

index	label	user_id
0	0	102156
1	1	1460

- offline test samples:

index	label	user_id
0	0	102404
1	1	1212

- online train samples:

index	label	user_id
0	0	102404
1	1	1212