企业竞价咨询问题分类与频次数据集
收藏资源简介：
1.数据采集：从电子卖场平台的帮助中心后台导出近2年所有用户咨询记录，筛选出栏目为“帮助中心”且问题文本中包含“竞价”“投标”“参与”“报名”等关键词的记录，采集字段包括：问题ID、栏目名称、提问主题、答复主题、问题原文、提问时间； 2.数据处理：1）清洗重复、无效及广告记录；规范化归并同一主体不同表述；2）用TF-IDF算法结合人工校验，提取“资质”“流程”“保证金”“报名时间”等高频关键词；将问题划分为资质类、流程类、技术类、其他类4类 3）计算每类问题提问量、占比、月度趋势；统计不同企业主体提问次数生成“高频提问企业名单”；按提问主体统计交互频次并生成高频主体清单；按答复主体统计服务效率指标并生成效率排序结果；4）形成含问题ID、分类标签、交互主体、时间、服务效率、统计指标的结构化数据集； 3.数据应用：该数据集用于优化帮助中心检索、输出投标企业问题清单、完善竞价指南培训，还可作为第三方数据产品输出行业咨询热度报告。
1. Data Collection: Export all user consultation records from the help center backend of the electronics mall platform over the past two years, filter records where the column is "Help Center" and the question text contains keywords such as "bidding", "tendering", "participate", "registration". The collected fields include: Question ID, Column Name, Question Subject, Reply Subject, Original Question Text, Question Time; 2. Data Processing: 1) Clean duplicate, invalid and advertising records; standardize and merge different expressions of the same subject; 2) Use TF-IDF algorithm combined with manual verification to extract high-frequency keywords such as "qualification", "process", "security deposit", "registration time"; classify the questions into four categories: Qualification Category, Process Category, Technical Category and Other Category; 3) Calculate the question volume, proportion and monthly trend of each category; count the number of inquiries from different enterprise subjects to generate a "High-frequency Inquiry Enterprise List"; generate a high-frequency subject list by counting interaction frequency based on the question subject; generate efficiency ranking results by counting service efficiency indicators based on the reply subject; 4) Form a structured dataset containing Question ID, classification label, interaction subject, time, service efficiency and statistical indicators; 3. Data Application: This dataset is used to optimize help center retrieval, output the inquiry list of bidding enterprises, improve bidding guide training, and can also serve as a third-party data product to output industry consultation heat reports.
