mvladimirova commited on
Commit
ee38d8e
ยท
verified ยท
1 Parent(s): c4cf1a6

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -20
README.md CHANGED
@@ -10,18 +10,6 @@ size_categories:
10
  - n>1T
11
  ---
12
 
13
- ---
14
- license: cc-by-nc-sa-4.0
15
- task_categories:
16
- - text-classification
17
- tags:
18
- - commerce
19
- - advertising
20
- - click
21
- size_categories:
22
- - n>1T
23
- ---
24
-
25
  # ๐Ÿ“Š Criteo 1TB Click Logs Dataset
26
 
27
  This dataset contains **feature values and click feedback** for millions of display ads. Its primary purpose is to **benchmark algorithms for clickthrough rate (CTR) prediction**.
@@ -29,7 +17,6 @@ This dataset contains **feature values and click feedback** for millions of disp
29
  It is similar, but larger than the dataset released for the Display Advertising Challenge hosted by Kaggle:
30
  ๐Ÿ”— [Kaggle Criteo Display Advertising Challenge](https://www.kaggle.com/c/criteo-display-ad-challenge)
31
 
32
- ---
33
 
34
  ## ๐Ÿ“ Full Description
35
 
@@ -42,8 +29,6 @@ This dataset contains **24 files**, each corresponding to **one day of data**.
42
  - The **first column** indicates whether the ad was **clicked (1)** or **not clicked (0)**.
43
  - Both **positive (clicked)** and **negative (non-clicked)** examples have been **subsampled**, though at **different rates** to keep business confidentiality.
44
 
45
- ---
46
-
47
  ## ๐Ÿงฑ Features
48
 
49
  - **13 integer features**
@@ -55,16 +40,14 @@ This dataset contains **24 files**, each corresponding to **one day of data**.
55
 
56
  > โš ๏ธ Some features may contain **missing values**.
57
 
58
- ---
59
 
60
  ## ๐Ÿงพ Data Format
61
 
62
  - Rows are **chronologically ordered**
63
  - Columns are **tab-separated** and follow this schema:
64
- <label> <int_feature_1> ... <int_feature_13> <cat_feature_1> ... <cat_feature_26>
65
  - If a value is missing, the field is simply **left empty**.
66
 
67
- ---
68
 
69
  ## ๐Ÿ”„ Differences from Kaggle Challenge Dataset
70
 
@@ -72,5 +55,4 @@ This dataset contains **24 files**, each corresponding to **one day of data**.
72
  - ๐Ÿ”„ **Subsampling ratios** differ
73
  - ๐Ÿ”ข **Ordering of features** is different
74
  - ๐Ÿงฎ Some features have **different computation methods**
75
- - ๐Ÿ” **Hash function** for categorical features has changed
76
-
 
10
  - n>1T
11
  ---
12
 
 
 
 
 
 
 
 
 
 
 
 
 
13
  # ๐Ÿ“Š Criteo 1TB Click Logs Dataset
14
 
15
  This dataset contains **feature values and click feedback** for millions of display ads. Its primary purpose is to **benchmark algorithms for clickthrough rate (CTR) prediction**.
 
17
  It is similar, but larger than the dataset released for the Display Advertising Challenge hosted by Kaggle:
18
  ๐Ÿ”— [Kaggle Criteo Display Advertising Challenge](https://www.kaggle.com/c/criteo-display-ad-challenge)
19
 
 
20
 
21
  ## ๐Ÿ“ Full Description
22
 
 
29
  - The **first column** indicates whether the ad was **clicked (1)** or **not clicked (0)**.
30
  - Both **positive (clicked)** and **negative (non-clicked)** examples have been **subsampled**, though at **different rates** to keep business confidentiality.
31
 
 
 
32
  ## ๐Ÿงฑ Features
33
 
34
  - **13 integer features**
 
40
 
41
  > โš ๏ธ Some features may contain **missing values**.
42
 
 
43
 
44
  ## ๐Ÿงพ Data Format
45
 
46
  - Rows are **chronologically ordered**
47
  - Columns are **tab-separated** and follow this schema:
48
+ > '''<label> <int_feature_1> ... <int_feature_13> <cat_feature_1> ... <cat_feature_26>'''
49
  - If a value is missing, the field is simply **left empty**.
50
 
 
51
 
52
  ## ๐Ÿ”„ Differences from Kaggle Challenge Dataset
53
 
 
55
  - ๐Ÿ”„ **Subsampling ratios** differ
56
  - ๐Ÿ”ข **Ordering of features** is different
57
  - ๐Ÿงฎ Some features have **different computation methods**
58
+ - ๐Ÿ” **Hash function** for categorical features has changed