服饰搭配点击率预估数据集

搭配简洁的配饰，如运动手表，方便监测心率和锻炼数据。 #生活知识# #健康生活# #健身技巧# #健身服装搭配#

The dataset of the paper Combo-Fashion: Fashion Clothes Matching CTR Prediction with Item History. This dataset is provided by Alibaba.

Overview

We collect users’ action records from two scenarios of Taobao to build these two datasets, termed as In-Shop Combo Dataset and Cross-Shop Combo Dataset, respectively. In-Shop Combo Dataset is collected from an in-shop matching scenario, where the top and bottom items are from the same store. Cross-Shop Combo Dataset is collected from another matching scenario where the top item and the bottom item can be from different stores. The pictures of combo items are synthesized based on the top and bottom item pictures, and are then displayed to customers. Some examples are presented in Figure 1.

Figure 1: An example of Combo-item

Dataset Description

This dataset contains three files and we have anonymized business-sensitive information. Considering the user privacy, this dataset does not public user-side features, which is a little different from the data used in our paper.

In-Shop Combo Dataset(in_shop_dataset.zip)

Sample Feature Set
There are 5 features in this set such as user_id, content_id(represent the Combo-Item ID), label(click or non-click), server_time and day. The "day" field is used to split the training set and testing set. For example, we can select the latest week, "day 33-39" for training and "day 40" for testing. Combo Item feature Set
For the features of the combo item feature side, we use "content__" as a prefix while the suffixes "_1" and "_2" represent the features of the items which make up the combo item respectively. It can be judged whether the item is top or bottom through the "content__part_1" and "content__part_2". The combo item feature set includes a total of 14 categorical form features and 3 categorical form features. Specially, "content__cpm_1" and "content__cpm_2" are the structured multi-hot feature which describe the item from three perspectives, namely characteristic, pattern (color & graphic design) and material.

Cross-Shop Combo Dataset(cross_shop_dataset.zip)
The data organization form is the same as the In-Shop Combo Dataset.

Pre-trained Item Embedding(pre_trained_item_embedding.zip)
Each row contains a 64-dimension item embedding feature which is trained based on our full traffic data.

Dataset Statistics

Dataset Users Items Combo Items Instances In-Shop Combo Dataset 2,639,413 212,726 5,271,278 35,605,809 Cross-Shop Combo Dataset 428,197 193,724 3,488,292 6,006,748

Citation

Please cite our paper if you use this dataset.

Chenxu Zhu, Peng Du, Weinan Zhang, Yong Yu, Yang Cao. 2022. Combo-Fashion: Fashion Clothes Matching CTR Prediction with Item History. In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining.

License

The dataset is distributed under the CC BY-SA 4.0 license.

Recommend Dataset

Fashion Clothes Matching CTR Prediction Dataset：https://tianchi.aliyun.com/dataset/dataDetail?dataId=131519

FashionAI - Attributes Recognition of Apparel Dataset：https://tianchi.aliyun.com/dataset/dataDetail?dataId=136923

FashionAI - Key Points Detection of Apparel Dataset: https://tianchi.aliyun.com/dataset/dataDetail?dataId=136948

Clothes Matching Dataset on Taobao：https://tianchi.aliyun.com/dataset/dataDetail?dataId=52

Catalog

Overview

Dataset Description

Dataset Statistics

Citation

License

Recommend Dataset

网址：服饰搭配点击率预估数据集 https://m.klqsh.com/news/view/156864

相关内容

时尚服饰搭配
如何轻松搭配时尚服饰？
服饰搭配论文(精选5篇)
服装色系搭配法3之搭配建议
各种颜色的衣服怎么搭配好看？服饰颜色搭配指南
服装搭配礼仪
时尚搭配技巧：分享实用的服装搭配建议，帮助人们提升穿搭水平
京东服饰携手小红书、Vogue服饰与美容解读2024秋冬四大穿搭新趋势
服装搭配AI应用
各季节服装搭配.ppt

随便看看