Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,76 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: cc-by-nc-sa-4.0
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-nc-sa-4.0
|
| 3 |
+
task_categories:
|
| 4 |
+
- text-classification
|
| 5 |
+
tags:
|
| 6 |
+
- click
|
| 7 |
+
- advertising
|
| 8 |
+
- commerce
|
| 9 |
+
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 |
+
# 📊 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**.
|
| 28 |
+
|
| 29 |
+
It is similar, but larger than the dataset released for the Display Advertising Challenge hosted by Kaggle:
|
| 30 |
+
🔗 [Kaggle Criteo Display Advertising Challenge](https://www.kaggle.com/c/criteo-display-ad-challenge)
|
| 31 |
+
|
| 32 |
+
---
|
| 33 |
+
|
| 34 |
+
## 📁 Full Description
|
| 35 |
+
|
| 36 |
+
This dataset contains **24 files**, each corresponding to **one day of data**.
|
| 37 |
+
|
| 38 |
+
### 🏗️ Dataset Construction
|
| 39 |
+
|
| 40 |
+
- The training data spans **24 days** of Criteo traffic.
|
| 41 |
+
- Each row represents a **display ad** served by Criteo.
|
| 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 |
+
## 🧱 Features
|
| 48 |
+
|
| 49 |
+
- **13 integer features**
|
| 50 |
+
Mostly count-based; represent numerical properties of the ad, user, or context.
|
| 51 |
+
|
| 52 |
+
- **26 categorical features**
|
| 53 |
+
Values are **hashed into 32-bit integers** for anonymization.
|
| 54 |
+
The **semantic meaning** of these features is **undisclosed**.
|
| 55 |
+
|
| 56 |
+
> ⚠️ Some features may contain **missing values**.
|
| 57 |
+
|
| 58 |
+
---
|
| 59 |
+
|
| 60 |
+
## 🧾 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 |
+
## 🔄 Differences from Kaggle Challenge Dataset
|
| 70 |
+
|
| 71 |
+
- 📅 The data covers a **different time period**
|
| 72 |
+
- 🔄 **Subsampling ratios** differ
|
| 73 |
+
- 🔢 **Ordering of features** is different
|
| 74 |
+
- 🧮 Some features have **different computation methods**
|
| 75 |
+
- 🔐 **Hash function** for categorical features has changed
|
| 76 |
+
|