SML310 Final Project Code and Slides
There are a few other approaches for providing recommendations beyond the traditional collaborative and content-based filtering architectures for recommendation systems. Some have received increasingly growing attention in research and industry, while others that remain relatively unexplored. Examples of techniques that fall in the former category are those that involve Deep Learning (usually relying upon neural networks), and have proved a helpful tool to improve upon the short-comings of the more traditional recommender system techniques. Meanwhile, an example of a technique that falls in the latter category would be the recent experimental efforts to use the stochastic process known as random walks to provide recommendations. In this paper, I wish to first explore the performance of basic content-based and collaborative recommendation systems on MovieLens 100k dataset, highlight their advantages and disadvantages, and compare them with a neural network-based recommendation system.
- ml-100k is the MovieLens 100k dataset used for this project downloaded from the GroupLens website.
- NN_Based_model_FormattedOutput is a GOOGLE colab notebook with the code for training, evaluating and predicting movie recommendations based on a neural network printing formatted recommendations
- SML310_proj python notebook for collaborative filtering and content based movie recommenders and then producing respective movie recommendations.
- SML310 Final presentation for end of term project.
All the listed resources were used in conception of this work. All figures used in the paper were created by author. Albayrak, S., & Bakır, Ç. (n.d.). User based and item based collaborative filtering with temporal dynamics. Retrieved December 02, 2020, from https://ieeexplore.ieee.org/abstract/document/6830213/ Collaborative Filtering Advantages & Disadvantages. (n.d.). Retrieved December 01, 2020, from https://developers.google.com/machine-learning/recommendation/collaborative/summary Collaborative Filtering | Recommendation Systems | Google Developers. (n.d.). Retrieved December 02, 2020, from https://developers.google.com/machine-learning/recommendation/collaborative/basics Content-based Filtering Advantages & Disadvantages. (n.d.). Retrieved December 01, 2020, from https://developers.google.com/machine-learning/recommendation/content-based/summary Content-based Filtering | Recommendation Systems | Google Developers. (n.d.). Retrieved December 01, 2020, from https://developers.google.com/machine-learning/recommendation/content-based/basics Isinkaye, F., Folajimi, Y., & Ojokoh, B. (2015, August 20). Recommendation systems: Principles, methods and evaluation. Retrieved December 01, 2020, from https://www.sciencedirect.com/science/article/pii/S1110866515000341 Kingma, D., & Ba, J. (2017, January 30). Adam: A Method for Stochastic Optimization. Retrieved December 01, 2020, from https://arxiv.org/abs/1412.6980 Le, Q., & Mikolov, T. (2014, May 22). Distributed Representations of Sentences and Documents. Retrieved December 01, 2020, from https://arxiv.org/abs/1405.4053 Meinl, R. (2020, May 13). Recommender Systems: The Most Valuable Application of Machine Learning (Part 1). Retrieved December 01, 2020, from https://towardsdatascience.com/recommender-systems-the-most-valuable-application-of-machine-learning-part-1-f96ecbc4b7f5 MovieLens 100K Dataset. (2019, May 20). Retrieved December 02, 2020, from https://grouplens.org/datasets/movielens/100k/ MovieLens. (2020, May 21). Retrieved December 01, 2020, from https://grouplens.org/datasets/movielens/ Rudin, C., & Radin, J. (2019, November 22). Why Are We Using Black Box Models in AI When We Don't Need To? A Lesson From An Explainable AI Competition · 1.2. Retrieved December 01, 2020, from https://hdsr.mitpress.mit.edu/pub/f9kuryi8/release/6 Sarwar, B. M. (n.d.). Challenges of User-based Collaborative Filtering Algorithms. Retrieved December 01, 2020, from http://www10.org/cdrom/papers/519/node9.html Semage, L. (2017, November 11). Recommender Systems with Random Walks: A Survey. Retrieved December 02, 2020, from https://arxiv.org/abs/1711.04101 Shani, G., & Gunawardana, A. (1970, January 01). Evaluating Recommendation Systems. Retrieved December 01, 2020, from https://link.springer.com/chapter/10.1007/978-0-387-85820-3_8 SPSS Tutorials: Pearson Correlation. (n.d.). Retrieved December 01, 2020, from https://libguides.library.kent.edu/SPSS/PearsonCorr Team, K. (n.d.). Keras documentation: Collaborative Filtering for Movie Recommendations. Retrieved December 01, 2020, from https://keras.io/examples/structured_data/collaborative_filtering_movielens/ What are Neural Networks? (n.d.). Retrieved December 01, 2020, from https://www.ibm.com/cloud/learn/neural-networks Wittenauer, J. (2019, April 29). Deep Learning With Keras: Recommender Systems. Retrieved December 01, 2020, from https://medium.com/@jdwittenauer/deep-learning-with-keras-recommender-systems-e7b99cb29929 Wu, W., He, L., & Yang, J. (n.d.). Evaluating Recommender Systems. Retrieved December 01, 2020, from https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=6360092 Zhang, S., Yao, L., Sun, A., & Tay, Y. (2019, July 10). Deep Learning based Recommender System: A Survey and New Perspectives. Retrieved December 01, 2020, from https://arxiv.org/abs/1707.07435 Zhang, S., Yao, L., Sun, A., & Tay, Y. (2019, July 10). Deep Learning based Recommender System: A Survey and New Perspectives. Retrieved December 02, 2020, from https://arxiv.org/abs/1707.07435