Data

Click_prediction_small

active
ARFF
Publicly available Visibility: public Uploaded 03-12-2014 by Jan van Rijn

1 likes downloaded by 2 people , 2 total downloads 0 issues 0 downvotes

1 likes downloaded by 2 people , 2 total downloads 0 issues 0 downvotes

Issue | #Downvotes for this reason | By |
---|

Loading wiki

Help us complete this description
Edit

Author: Tencent Inc.
Source: [KDD Cup](https://www.kddcup2012.org/) - 2012
Please cite:
This data set is the same as version 4, but has additional unlabeled data attached to it. This is meant for a machine learning challenge. The complete labeled version of this dataset is version 6 (but this version is kept private for the duration of the challenge).
This data is derived from the 2012 KDD Cup. The data is subsampled to 1% of the original number of instances, downsampling the majority class (click=0) so that the target feature is reasonably balanced (5 to 1).
The data is about advertisements shown alongside search results in a search engine, and whether or not people clicked on these ads.
The task is to build the best possible model to predict whether a user will click on a given ad.
A search session contains information on user id, the query issued by the user, ads displayed to the user, and target feature indicating whether a user clicked at least one of the ads in this session. The number of ads displayed to a user in a session is called ‘depth’. The order of an ad in the displayed list is called ‘position’. An ad is displayed as a short text called ‘title’, followed by a slightly longer text called ’description’, and a URL called ‘display URL’.
To construct this dataset each session was split into multiple instances. Each instance describes an ad displayed under a certain setting (‘depth’, ‘position’). Instances with the same user id, ad id, query, and setting are merged. Each ad and each user have some additional properties located in separate data files that can be looked up using ids in the instances.
The dataset has the following features:
* Click – binary variable indicating whether a user clicked on at least one ad.
* Impression - the number of search sessions in which AdID was impressed by UserID who issued Query.
* Url_hash - URL is hashed for anonymity
* AdID
* AdvertiserID - some advertisers consistently optimize their ads, so the title and description of their ads are more attractive than those of others’ ads.
* Depth - number of ads displayed to a user in a session
* Position - order of an ad in the displayed list
* QueryID - is the key of the data file 'queryid_tokensid.txt'. (follow the link to the original KDD Cup page, track 2)
* KeywordID - is the key of 'purchasedkeyword_tokensid.txt' (follow the link to the original KDD Cup page, track 2)
* TitleID - is the key of 'titleid_tokensid.txt'
* DescriptionID - is the key of 'descriptionid_tokensid.txt' (follow the link to the original KDD Cup page, track 2)
* UserID – is also the key of 'userid_profile.txt' (follow the link to the original KDD Cup page, track 2). 0 is a special value denoting that the user could be identified.

click (target) | nominal | 2 unique values 399482 missing | |

impression | numeric | 327 unique values 0 missing | |

url_hash (ignore) | numeric | 15787 unique values 0 missing | |

ad_id | numeric | 82778 unique values 0 missing | |

advertiser_id | numeric | 11955 unique values 0 missing | |

depth | numeric | 3 unique values 0 missing | |

position | numeric | 3 unique values 0 missing | |

query_id (ignore) | numeric | 248486 unique values 0 missing | |

keyword_id | numeric | 93414 unique values 0 missing | |

title_id | numeric | 155012 unique values 0 missing | |

description_id | numeric | 129271 unique values 0 missing | |

user_id | numeric | 287967 unique values 0 missing |

100889.02

Second quartile (Median) of standard deviation of attributes of the numeric type.

0.52

Area Under the ROC Curve achieved by the landmarker weka.classifiers.trees.RandomTree -depth 1

Maximum mutual information between the nominal attributes and the target attribute.

2

The minimal number of distinct values among attributes of the nominal type.

0.25

Error rate achieved by the landmarker weka.classifiers.trees.RandomTree -depth 1

Number of attributes needed to optimally describe the class (under the assumption of independence among attributes). Equals ClassEntropy divided by MeanMutualInformation.

2

The maximum number of distinct values among attributes of the nominal type.

38.16

Third quartile of kurtosis among attributes of the numeric type.

0.12

Kappa coefficient achieved by the landmarker weka.classifiers.trees.RandomTree -depth 1

0.54

Area Under the ROC Curve achieved by the landmarker weka.classifiers.trees.J48 -C .00001

1930301.79

Third quartile of means among attributes of the numeric type.

0.56

Area Under the ROC Curve achieved by the landmarker weka.classifiers.trees.DecisionStump -E "weka.attributeSelection.CfsSubsetEval -P 1 -E 1" -S "weka.attributeSelection.BestFirst -D 1 -N 5" -W

0.52

Area Under the ROC Curve achieved by the landmarker weka.classifiers.trees.RandomTree -depth 2

Third quartile of mutual information between the nominal attributes and the target attribute.

0.16

Error rate achieved by the landmarker weka.classifiers.trees.DecisionStump -E "weka.attributeSelection.CfsSubsetEval -P 1 -E 1" -S "weka.attributeSelection.BestFirst -D 1 -N 5" -W

0.25

Error rate achieved by the landmarker weka.classifiers.trees.RandomTree -depth 2

0.07

Kappa coefficient achieved by the landmarker weka.classifiers.trees.J48 -C .00001

0.54

Area Under the ROC Curve achieved by the landmarker weka.classifiers.bayes.NaiveBayes

5.52

Third quartile of skewness among attributes of the numeric type.

0.08

Kappa coefficient achieved by the landmarker weka.classifiers.trees.DecisionStump -E "weka.attributeSelection.CfsSubsetEval -P 1 -E 1" -S "weka.attributeSelection.BestFirst -D 1 -N 5" -W

0.12

Kappa coefficient achieved by the landmarker weka.classifiers.trees.RandomTree -depth 2

0.54

Area Under the ROC Curve achieved by the landmarker weka.classifiers.trees.J48 -C .0001

-0.93

First quartile of kurtosis among attributes of the numeric type.

2979990.45

Third quartile of standard deviation of attributes of the numeric type.

0.56

Area Under the ROC Curve achieved by the landmarker weka.classifiers.bayes.NaiveBayes -E "weka.attributeSelection.CfsSubsetEval -P 1 -E 1" -S "weka.attributeSelection.BestFirst -D 1 -N 5" -W

0.52

Area Under the ROC Curve achieved by the landmarker weka.classifiers.trees.RandomTree -depth 3

0.57

Area Under the ROC Curve achieved by the landmarker weka.classifiers.trees.REPTree -L 1

0.16

Error rate achieved by the landmarker weka.classifiers.bayes.NaiveBayes -E "weka.attributeSelection.CfsSubsetEval -P 1 -E 1" -S "weka.attributeSelection.BestFirst -D 1 -N 5" -W

0.25

Error rate achieved by the landmarker weka.classifiers.trees.RandomTree -depth 3

0.07

Kappa coefficient achieved by the landmarker weka.classifiers.trees.J48 -C .0001

Average mutual information between the nominal attributes and the target attribute.

First quartile of mutual information between the nominal attributes and the target attribute.

0.08

Kappa coefficient achieved by the landmarker weka.classifiers.bayes.NaiveBayes -E "weka.attributeSelection.CfsSubsetEval -P 1 -E 1" -S "weka.attributeSelection.BestFirst -D 1 -N 5" -W

0.12

Kappa coefficient achieved by the landmarker weka.classifiers.trees.RandomTree -depth 3

0.54

Area Under the ROC Curve achieved by the landmarker weka.classifiers.trees.J48 -C .001

An estimate of the amount of irrelevant information in the attributes regarding the class. Equals (MeanAttributeEntropy - MeanMutualInformation) divided by MeanMutualInformation.

-0.28

First quartile of skewness among attributes of the numeric type.

0.1

Kappa coefficient achieved by the landmarker weka.classifiers.trees.REPTree -L 1

0.56

Area Under the ROC Curve achieved by the landmarker weka.classifiers.lazy.IBk -E "weka.attributeSelection.CfsSubsetEval -P 1 -E 1" -S "weka.attributeSelection.BestFirst -D 1 -N 5" -W

0

Standard deviation of the number of distinct values among attributes of the nominal type.

2

Average number of distinct values among the attributes of the nominal type.

17.37

First quartile of standard deviation of attributes of the numeric type.

0.57

Area Under the ROC Curve achieved by the landmarker weka.classifiers.trees.REPTree -L 2

0.16

Error rate achieved by the landmarker weka.classifiers.lazy.IBk -E "weka.attributeSelection.CfsSubsetEval -P 1 -E 1" -S "weka.attributeSelection.BestFirst -D 1 -N 5" -W

0.08

Kappa coefficient achieved by the landmarker weka.classifiers.lazy.IBk -E "weka.attributeSelection.CfsSubsetEval -P 1 -E 1" -S "weka.attributeSelection.BestFirst -D 1 -N 5" -W

2.33

Second quartile (Median) of kurtosis among attributes of the numeric type.

0.1

Kappa coefficient achieved by the landmarker weka.classifiers.trees.REPTree -L 2

35086.02

Second quartile (Median) of means among attributes of the numeric type.

0.57

Area Under the ROC Curve achieved by the landmarker weka.classifiers.trees.REPTree -L 3

0.53

Area Under the ROC Curve achieved by the landmarker weka.classifiers.trees.DecisionStump

Second quartile (Median) of mutual information between the nominal attributes and the target attribute.

0.17

Error rate achieved by the landmarker weka.classifiers.trees.DecisionStump

1.76

Second quartile (Median) of skewness among attributes of the numeric type.

0.1

Kappa coefficient achieved by the landmarker weka.classifiers.trees.REPTree -L 3

0

Kappa coefficient achieved by the landmarker weka.classifiers.trees.DecisionStump

Minimal mutual information between the nominal attributes and the target attribute.