A generative adversarial network–based method for generating negative financial samples

Abstract

In financial anti-fraud field, negative samples are small and sparse with serious sample imbalanced problem. Generating negative samples consistent with original data to naturally solve imbalanced problem is a serious problem. This article proposes a new method to solve this problem. We introduce a new generation model, combined Generative Adversarial Network with Long Short-Term Memory network for one-dimensional negative financial samples. The characteristic association between transaction sequences can be learned by long short-term memory layer, and the generator covers real data distribution by the adversarial discriminator with time-sequence. Mapping data distribution to feature space is a common evaluation method of synthetic data; however, relationships between data attributes have been ignored in online transactions. We define a comprehensive evaluation method to evaluate the validity of generated samples from data distribution and attribute characteristics. Experimental results on real bank B2B transaction data show that the proposed model has higher overall ratings, which is 10% higher than traditional generation models. Finally, well-trained model is used to generate negative samples and form new dataset. The classification results on new datasets show that precision and recall are all higher than baseline models. Our work has a certain practical value and provides a new idea to solve imbalanced problem in whatever fields.

Gradually Generative Adversarial Networks Method for Imbalanced Datasets

“…Meanwhile, the NB method is used for classification. The test results have resulted in an accuracy value of 88.5%, more significant than the R algorithm of 87.5% [23]. To overcome the imbalance problem, the popular SMOTE algorithm is used while the k-NN method is chosen for classification.…”

“…For example, in the field of financial antifraud with smaller data samples, the GAN model can be collaborated with the Long Short-Term Memory (LSTM) network algorithm so that the problem of data imbalance can be taken seriously [28]. Both of these models will share roles to process data in a time sequence completed by the Long Short-Term Memory network, while the GAN is to distribute selected real data which will produce data that is similar to the original data [23]. In this research, the SMOTE algorithm was used to overcome data imbalance, while the best method was chosen to classify, namely NB, D-Tree and RF, for example in the case of breast disease prediction [24].…”

“…In this paper, to handle data imbalance, the SMOTE and Borderline-SMOTE algorithms were selected, while to measure the accuracy value, they relied on the Safe Level Graph [13]. In the end, the GAN model will produce data similar to the original [23]. In this paper, to overcome data imbalances, the SMOTE algorithm is used, while for classification, the method chosen is D-Tree [26].…”

“…NB has the advantage that it is simple and works well in the real world. The weakness is that the dataset to be tested must be in numerical form [22], [23]. K-NN has advantages, which are effective and robust against noisy training data [24].…”

Fraud Detection System Using Facial Recognition Based On Google Teachable Machine for Banking Applications

“…Most of the time, interested parties are reserved to provide their information and it becomes difficult to detect whether the information is being provided to the real or correct or false person, so the privacy related to personal information is subject to significant challenges and other challenges are also important. Zhang, Yang, Chen, Liu, Meng, Wang and Li (2020) [13], in their research work, they proposed new techniques to overcome financial risk or fraud by generating negative financial samples through the method based on conflicting generative networks. This method helps to solve the unbalanced dataset problem.…”