Wine Reviews 葡萄酒评论

在看了Somm(一个关于侍酒师的纪录片)之后，我想知道如何像侍酒师那样通过盲品来创建一个预测模型来鉴别葡萄酒。此过程的第一步是收集一些数据来训练模型。我计划使用深度学习通过描述/评论中的文字来预测葡萄酒的种类。该模型仍然不能品尝葡萄酒，但理论上它可以根据侍酒师给出的描述识别葡萄酒。我的总体目标是创建一个模型，它可以根据描述识别葡萄酒的品种、酿酒厂和位置。

After watching Somm, a documentary about sommeliers, I aimed to investigate how to develop a predictive model for wine identification via blind tasting, just as professional sommeliers do. The first step of this workflow is to collect relevant datasets for model training. I plan to employ deep learning techniques to predict wine varieties using textual content from wine descriptions or reviews. While this model cannot actually taste wine, it can theoretically identify wines based on the descriptions provided by sommeliers. My overall objective is to build a model that can recognize a wine's variety, winery, and production location according to its associated descriptive text.

搜集汇总

背景与挑战

背景概述

该数据集名为'Wine Reviews 葡萄酒评论'，包含约13万至15万条葡萄酒评论数据，涵盖品种、位置、酒庄、价格和描述等字段。数据源自WineEnthusiast网站，旨在支持通过文本描述预测葡萄酒属性的模型训练。