One-Class Adversarial Nets for Fraud Detection
Authors: Panpan Zheng, Shuhan Yuan, Xintao Wu, Jun Li, Aidong Lu1286-1293
AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experimental results show that our OCAN outperforms the state-of-the-art oneclass classification models and achieves comparable performance with the latest multi-source LSTM model that requires both benign and malicious users in the training phase. |
| Researcher Affiliation | Academia | 1University of Arkansas, {pzheng,sy005,xintaowu}@uark.edu 2University of Oregon, lijun@cs.uoregon.edu 3University of North Carolina at Charlotte, aidong.lu@uncc.edu |
| Pseudocode | Yes | Algorithm 1: Training One-Class Adversarial Nets |
| Open Source Code | Yes | Our software together with the datasets are available at https://github.com/Panpan Zheng/OCAN. |
| Open Datasets | Yes | To evaluate OCAN, we focus on one type of malicious users, i.e., vandals on Wikipedia. We conduct our evaluation on UMDWikipedia dataset (Kumar, Spezzano, and Subrahmanian 2015). |
| Dataset Splits | Yes | Note that both OCNN and OCGP require a small portion (5% in our experiments) of vandals as a validation dataset to tune an appropriate threshold for vandal detection. However, OCAN does not require any vandals for training and validation. To evaluate the performance of vandal detection, we randomly select 7000 benign users as the training dataset and 3000 benign users and 3000 vandals as the testing dataset. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., exact GPU/CPU models, processor types with speeds, memory amounts) used for running its experiments. |
| Software Dependencies | No | The paper mentions software like TensorFlow in citations, but does not provide specific version numbers for its own software dependencies (e.g., library or solver names with versions). |
| Experiment Setup | Yes | For LSTM-Autoencoder, the dimension of the hidden layer is 200, and the training epoch is 20. For the complementary GAN model, both discriminator and generator are feedforward neural networks. Specifically, the discriminator contains 2 hidden layers which are 100 and 50 dimensions. The generator takes the 50 dimensions of noise as input, and there is one hidden layer with 100 dimensions. The output layer of the generator has the same dimension as the user representation which is 200 in our experiments. The training epoch of complementary GAN is 50. The threshold ϵ defined in Equation 12 is set as the 5-quantile probability of real benign users predicted by a pre-trained discriminator. |