Amazon Fine Food Reviews dataset
收藏资源简介:
该数据集包含亚马逊上精细食品的评论,用于构建情感分析器。数据集详细记录了每条评论的ID、产品ID、用户ID、用户名、评分、帮助性评分分子和分母、时间、总结和文本内容。
This dataset comprises reviews of fine foods from Amazon, intended for the construction of sentiment analyzers. It meticulously documents each review's ID, product ID, user ID, username, rating, helpfulness score numerator and denominator, timestamp, summary, and text content.
数据集概述
数据集名称
- Sentiment Analyzer
数据集类型
- NLP (Natural Language Processing)
数据集领域
- Machine Learning
数据集目的
- 构建一个先进的情感分析器,用于分析亚马逊精细食品评论。
数据集结构
-
原始数据结构
| id | ProductId | UserId | ProfileName | Score | HelpfulnessNumerator | HelpfulnessDenominator | Time | Summary | Text |
示例:
| 1 | B001E4KFG0 | A3SGXH7AUHU8GW | delmartian | 5 | 1 | 1 | 1303862400 | Good Quality Dog Food | I have bought several of the Vitality... |
-
预处理后数据结构
| Text | Sentiment |
示例:
| I have to say I was a little apprehensive to b... | 1 | | Received my free K cups as a sample promotion ... | 1 | | Brooklyn "French Roast" K-Cup Coffee is not on... | 0 |
数据预处理
- 包括数据清洗和特征工程,以使数据适合机器学习模型。
模型与性能
- Bag of Words (BOW)
- 性能:90.99%
- Bag of Words with Stemming
- 性能:90.51%
- Bag of Words with Lemmatization
- 性能:90.86%
- Bag of Words with n-grams (Bi-gram)
- 性能:91.13%
- Tri-gram性能:87.35%
- Combined Approach (Bag of words with Lemmatization and Bi-gram features)
- 性能:91.21%
- Binary Bag of Words with lemmatization and Bi-gram features性能:90.82%
- Term Frequency Inverse Document frequency (TFIDF)
- 性能:85.13%
- Average Word2Vec
- 性能:91.14%
- Recurrent Neural Network with Word2Vec
- 性能:95.26%
依赖库
- pandas, numpy, sklearn, matplotlib, seaborn, tensorflow, pytorch, nltk, spacy




