京津冀地区工商业用户安全等级评估数据

工商业用户由于用气量大，作业环境复杂，周边环境人流量大，往往是燃气安全的重点监管场所。采集工商业用户的每日用气数据，用气设备数量和信息等，运用大数据构建其正常用气行为特征，得到正常用气时间范围、用气量值范围等信息，对其安全用气行为进行预测，如发现异常，可以及时提示报警。1、获取用户的历史数据，包括地区、户号、每日的用气量、设备编号等维度特征，运用大数据分析其历史行为，得到最大用气量、最小用气量、平均用气量等数据信息；2、通过用户开户日期、安装日期、用气时间，结合GBDT算法进行建模，形成用户画像；3、通过DeepFM深度学习算法，进行用气行为安全的预测，并结合历史数据，计算等级标签。

Commercial and industrial (C&I) users, characterized by high gas consumption, complex operating environments and high foot traffic in surrounding areas, are usually key supervision targets for gas safety. By collecting daily gas consumption data, quantity and information of gas-using equipment and other relevant information of C&I users, big data is employed to construct the characteristics of their normal gas consumption behavior, including the normal gas consumption time range and gas consumption value range, so as to predict their safe gas consumption behavior and issue timely alerts once abnormalities are detected. 1. Collect historical user data covering dimensional features such as region, household number, daily gas consumption and equipment number, analyze their historical behavior via big data, and obtain data metrics including maximum gas consumption, minimum gas consumption and average gas consumption; 2. Build models combined with the GBDT algorithm based on the user's account opening date, installation date and gas usage time to generate user portraits; 3. Use the DeepFM deep learning algorithm to predict the safety of gas consumption behavior, and calculate grade labels in combination with historical data.