银行客户交易场景智能分类

Intelligent Classification of Bank Customer Transaction Scene

如何聚合多维度海量数据,充分挖掘数据的内在价值,是商业银行数字化转型的重要目标之一。文章以银行个人客户支付交易数据为例,探索Textgrocery与FastText两类自然语言处理算法对交易所属场景的分类效果。实验表明,Textgrocery效果在分场景和整体方面均优于FastText,故文章最终选择Textgrocery算法对交易数据所属场景开展自动化、智能化、高效化分类。模型场景分类结果可以帮助银行为个人客户建立消费行为维度的客户标签,从而使数据资源变得可使用、有价值。

How to aggregate multi-dimensional massive data and fully tap the intrinsic value of data is one of the important objectives in commercial bank digital transformation.Taking the payment transaction data of individual customers as an example,the article explores the classification effect of two kinds of natural language processing algorithms,Textgrocery and FastText,on the classification of transaction data scenes.The experimental results show that the effect of Textgrocery is superior to FastText in both scene segmentation and overall performance,so the article finally applies Textgrocery algorithm on automatically,intelligently and efficiently classification of the scenes of transaction data.The results of scene classification model can help banks to establish labels for individual customers in the dimension of consumption behavior,thus making data resources usable and valuable.