electricsheepafrica/nigerian-banking-ussd-banking-logs

这是一个具有尼日利亚银行实际情境的合成数据集，包含USSD欺诈标签，旨在用于USSD欺诈检测、特征手机银行优化和可访问性分析。数据集具有尼日利亚语境，包括真实的银行名称、州、人口统计信息。数据集规模较大，包含200万个条目，覆盖了从2023年1月1日到2024年12月31日的时间段。数据集分为训练集、验证集和测试集，分别包含140万、30万和30万个条目。关键列包括会话ID、时间戳、客户ID、电话号码、USSD代码、菜单深度、会话时长（秒）、交易类型、会话状态和欺诈标志。欺诈标签的普及率为1.2%。

This is a synthetic dataset with authentic Nigerian banking context, containing USSD fraud labels, designed for USSD fraud detection, feature phone banking optimization, and accessibility analytics. The dataset includes Nigerian context with real bank names, states, and demographics. It is large-scale with 2,000,000 entries, covering the period from 2023-01-01 to 2024-12-31. The dataset is split into training (1,400,000 entries), validation (300,000 entries), and test (300,000 entries) sets. Key columns include session_id, timestamp, customer_id, phone_number, ussd_code, menu_depth, session_duration_seconds, transaction_type, session_status, and fraud_flag, with a fraud flag prevalence of 1.2%.