借贷意向衍生特征-趋势B

提出多切片变量衍生字段，基于不同时间切片的数据计算比例、均值、斜率等指标作为原始数据的补充，结合以周为维度的高频监控，能够更加精准地量化用户风险

Multi-slice derived variable fields are proposed, which calculate metrics such as ratios, means, and slopes using data from different time slices as supplements to the original dataset. Combined with weekly high-frequency monitoring, this approach enables more accurate quantification of user risks.