google/extended_amazon_2023_dataset

--- pretty_name: Extended Amazon 2023 Dataset source_dataset: https://amazon-reviews-2023.github.io/ language: - en license: cc-by-4.0 tags: - jsonl - text - recommendation - shopping - image descriptions - llm generated metadata size_categories: - 10M<n<100M dataset_info: features: - name: data_source dtype: string - name: product_id dtype: string - name: image_description dtype: string - name: title_to_display dtype: string - name: description_to_display dtype: string - name: derived_features sequence: string splits: - name: dataset num_examples: 9008280 --- # Amazon Reviews '23 Extension Dataset This dataset provides an extension of the Amazon Reviews ’23 Dataset ([https://amazon-reviews-2023.github.io/](https://amazon-reviews-2023.github.io/)). ## Categories Included * Appliances * Clothing_Shoes_and_Jewelry * Sports_and_Outdoors * Videos_Games ## Methodology First, we performed a series of cleaning and data augmentation steps. We removed the small fraction of items where there was not an image, as having an image is critical to answer our research questions. Another issue we found with the images was that all six items from a search (e.g., for shoes) would be the same item, say of different sizes. To address this, we treated all items that have the same image associated with them as a single image. Then, we used an LLM call where the prompt included the provided title, provided description (which often is not in the original data), features, and an LLM-generated description of the image itself to create the displayed user-friendly item title and item description. ## Data Schema The data is provided in **jsonl** format where each line is a JSON object with the following fields: | Field | Description | | :--------------------- | :------------------------------------------------------------------------------------------------------- | | `data_source` | The category from which the data came (e.g., "Clothing, Shoes & Jewelry"). | | `product_id` | The `product_id` from the original Amazon Review '23 dataset to enable joining with the original data. | | `image_description` | An LLM-generated description of the image. | | `title_to_display` | An LLM-generated title based upon the original title and the image description. | | `description_to_display` | An LLM-generated item description (based on the original metadata and the image description). | | `derived_features` | An LLM-generated set of features (based on the original metadata and the image description). | ## Sample Entry Below is one sample entry: ```json { "data_source": "Clothing, Shoes & Jewelry", "product_id": "B09CG2SPHX", "image_description": "These Crocs Capri Strappy Flip Flops boast a stylish blend of textures and patterns. The straps feature a lightweight, subtly shimmering leopard print fabric, offering a touch of playful sophistication. This is beautifully contrasted by the smooth, neutral-toned footbed and the subtly textured, dark brown, molded outsole providing a comfortable and durable base. The overall shape is sleek and modern, creating a comfortable yet fashionable sandal perfect for warm-weather wear.", "title_to_display": "Crocs Women's Capri Sandals", "description_to_display": "The Crocs Women's Capri Strappy Flip Flops are a stylish and comfortable option for warm-weather wear. These sandals feature soft, microfiber-lined straps with a stylish synthetic metallic pattern. Designed with a thicker, Dual Crocs Comfort footbed, a deep heel cradle, and a Croslite foam outsole, they cushion every step. With a blend of textures and patterns, including a subtly shimmering leopard print fabric, these flip flops offer a touch of playful sophistication. The smooth, neutral-toned footbed and textured, dark brown, molded outsole provide a comfortable and durable base. The lightweight design and relaxed fit make them perfect for dressing up or hanging out. Made with 100% synthetic materials and an Ethylene Vinyl Acetate sole, these imported sandals have a heel that measures approximately 1.57''.", "derived_features": [ "Synthetic", "Imported", "EVA sole", "1.57'' heel", "Dual Straps", "Comfortable", "Lightweight", "Durable", "Stylish", "Metallic Straps" ] }--- pretty_name: 扩展版亚马逊2023数据集 source_dataset: https://amazon-reviews-2023.github.io/ language: - 英语 license: 知识共享署名4.0（CC BY 4.0） tags: - jsonl - 文本 - 推荐系统 - 购物 - 图像描述 - 大语言模型（Large Language Model）生成元数据 size_categories: - 1000万 < 样本数 < 1亿 dataset_info: features: - name: data_source dtype: 字符串 - name: product_id dtype: 字符串 - name: image_description dtype: 字符串 - name: title_to_display dtype: 字符串 - name: description_to_display dtype: 字符串 - name: derived_features dtype: 字符串序列 splits: - name: dataset num_examples: 9008280 --- # 亚马逊评论'23扩展数据集 本数据集为亚马逊评论'23数据集（https://amazon-reviews-2023.github.io/）的扩展版本。 ## 包含类别 * 家电 * 服饰鞋靴与珠宝 * 运动户外 * 视频游戏 ## 处理方法 首先，我们执行了一系列数据清洗与增强步骤。我们移除了无图像的少量条目，因为图像是解答我们研究问题的关键要素。 我们发现的另一项图像相关问题是：单次搜索返回的6个商品（例如搜索鞋子）会为同一商品的不同尺寸展示相同图像。为解决该问题，我们将关联同一图像的所有条目视为同一图像单元。 随后，我们调用大语言模型，以提供的原始标题、原始描述（原始数据中常缺失该内容）、商品特征以及大语言模型生成的图像描述作为提示词，生成供用户查看的友好型商品标题与商品描述。 ## 数据格式 数据以**jsonl**格式提供，其中每行均为一个JSON对象，包含以下字段： | 字段名 | 字段说明 | | :--------------------- | :------------------------------------------------------------------------------------------------------- | | `data_source` | 数据所属的商品类别（例如："服饰鞋靴与珠宝"）。 | | `product_id` | 原始亚马逊评论'23数据集中的`product_id`，用于与原始数据集进行关联。 | | `image_description` | 大语言模型生成的图像描述。 | | `title_to_display` | 基于原始标题与图像描述生成的展示用商品标题。 | | `description_to_display` | 基于原始元数据与图像描述生成的展示用商品描述。 | | `derived_features` | 基于原始元数据与图像描述生成的一组商品特征。 | ## 示例条目 以下为一条示例条目： json { "data_source": "服饰鞋靴与珠宝", "product_id": "B09CG2SPHX", "image_description": "这款卡骆驰（Crocs）Capri系带人字拖采用别致的纹理与图案拼接设计。鞋带采用轻盈且带有微妙光泽的豹纹面料，增添了俏皮精致感。光滑的中性色调鞋床与带有细微纹理的深棕色模压外底形成了美观对比，提供了舒适且耐用的支撑。整体造型简约现代，打造出舒适又时尚的凉鞋，非常适合温暖天气穿着。", "title_to_display": "卡骆驰女士Capri凉鞋", "description_to_display": "卡骆驰女士Capri系带人字拖是温暖天气穿搭的时尚舒适之选。这款凉鞋采用柔软的微纤维衬里鞋带，搭配别致的合成金属质感图案。搭载更厚实的双舒适鞋床（Dual Crocs Comfort）、深足跟承托区以及Croslite泡沫外底，为每一步提供缓冲。结合多种纹理与图案，包括带有微妙光泽的豹纹面料，这款人字拖尽显俏皮精致感。光滑的中性色调鞋床与带有纹理的深棕色模压外底提供了舒适且耐用的支撑。轻盈的设计与宽松的版型使其既适合穿搭点缀，也适合日常休闲。这款进口凉鞋采用100%合成材料与乙烯-醋酸乙烯酯共聚物（EVA）外底，鞋跟高度约为1.57英寸。", "derived_features": [ "合成材料", "进口商品", "EVA外底", "1.57英寸鞋跟", "双鞋带设计", "舒适", "轻盈", "耐用", "时尚", "金属质感鞋带" ] }