Attention-Based Transactional Context Embedding for Next-Item Recommendation

Authors: Shoujin Wang, Liang Hu, Longbing Cao, Xiaoshui Huang, Defu Lian, Wei Liu

AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The empirical study on real-world transaction datasets proves that ATEM significantly outperforms the state-of-the-art methods in terms of both accuracy and novelty.
Researcher Affiliation Academia GBDTC, FEIT, Univeristy of Technology Sydney Advanced Analytics Institute, Univeristy of Technology Sydney *Big Data Research Center, University of Electronic Science and Technology of China
Pseudocode Yes Algorithm 1 ATEM Parameter Learning Using SGD
Open Source Code No The paper does not provide any explicit statements or links indicating that the source code for the described methodology is publicly available.
Open Datasets Yes We evaluate our method on two real-world transaction data sets: IJCAI-15 1 and Tafang 2. 1https://tianchi.aliyun.com/datalab/dataSet.htm?id=1 2http://stackoverflow.com/questions/25014904/download-linkfor-ta-feng-grocery-dataset
Dataset Splits Yes we randomly choose 20% from the transactions happened in last 30 days as the testing set, while the remainder is for training.
Hardware Specification No The paper does not specify any particular hardware (e.g., GPU models, CPU types) used for running the experiments.
Software Dependencies No The paper mentions using Adam for gradient descent but does not specify version numbers for any programming languages or software libraries used.
Experiment Setup Yes For our ATEM model, the batch size is empirically set to 50 and the number of hidden units for item embeddings is set to 128 and 40 on IJCAI-15 and Tafang dataset respectively. We run 20 epochs to train the model. K is empirically set to 8 in our experiments. The parameter λ in exponential decay is set to 0.75 to obtain the best performance.