COMP4135/7240 Group Assignment
系统开发代写 The purpose of this assignment is to assess how well students have learned the fundamentals of recommender system implementation…
Read carefully the instructions of this assignment.
The purpose of this assignment is to assess how well students have learned the fundamentals of recommender system implementation and evaluation, and know how to apply the technique in solving a real-world problem.
Tasks 系统开发代写
- Propose a recommender system that is able to generate personalized recommendations of a specific type of product on your choice (e.g., music, movies, books, courses, jokes, articles, etc.).
o Note: the choice of product type depends on the dataset you have.
- System development: the system should contain the following components/steps and can be run on a website or a mobile app.
- Preference elicitation: a user interface that is able to elicit users’ explicit preferences (e.g., ratings) on the items, based on which the recommendations can be generated.
- Core computation algorithms: implementation of two different recommendation methods (on your choice) that have been introduced in the class.
- Result presentation: a user interface that displays the generated recommendations to users. Each recommendation should be associated with an explanation that can explain to the user why that item is recommended to her/him.
- Feedback elicitation and recommendation refinement: the user interface should allow users to give feedback on the recommendation, in which case a new set of recommendations should be returned based on user feedback. 系统开发代写
- System evaluation: ask at least 20 persons (not classmates) to evaluate the accuracy of your system’s recommendations.
- Testing set-up: All users are randomly divided into two groups, and assigned to evaluate the recommendations by the two different methods you developed respectively. For example, if there are 20 users and you implement two recommendation methods A and B, 10 users are asked to evaluate A and the other 10 users are asked to evaluate B.
- Evaluation procedure: You may ask each user to rate every recommendation, and then compute the user’s average rating on all of the recommended items.
-
Results analysis:
- Within each method, you can analyze whether users’ ratings in the first round of recommendations are significantly different from those in the 2nd round of recommendations (after their feedback), through the Paired t-Test (see reference below).
- Comparison of the two methods: you can compare the two methods (e.g., A and B) for the first round of recommendations by conducting the Two-Sample t-Test (see reference below).
Deliverables 系统开发代写
- Report (in a single Word or PDF file)
- Cover page: group name, all members’ names and IDs.
- System description: (About 1-2 pages)
o Description of the recommender system that your group proposes and its main features.
o Description of the two different recommendation methods that your group chooses and the main reasons for choosing them.
- Screenshots of all user interfaces (each with one-sentence description).
- Evaluation results that should include information of users you invite to do the testing, the evaluation procedure, and analysis results.
- Reflection by each group member (individual work):
o Your role in the whole process of system development and evaluation. (About 1 paragraph)
o Your observations of any issues with the current system implementation, i.e., regarding its user interface design and choice of recommendation methods, and your suggestions for further enhancements. (About 1-2 paragraphs)
o Note: this part will be individually assessed, so clearly indicate the member’s ID and name.
- References (such as the datasets and development toolkits you have used).
- Signed group participation form (see the attached) in the scanned copy.
o Note: the assignment will NOT be graded if the group does not submit the hand-signed participant form by members.
- Source files (in a .zip file)
- All files needed to run your system.
- A read me file to instruct how to run.
- A URL for online testing (optional).
-
Video presentation (10-min in MP4 or MOV format) 系统开发代写
- Your group gives a video presentation about this project, which should include the system description, user interface design, implemented recommendation methods, and evaluation results.
- Every member should present for at least 3 minutes.
- The video should display clearly the presenter’s face at the right up or bottom corner of the video that allows the lecturer to be able to identify the student’s identity during the presentation. Assignment submission
- Let one member of your group submit the assignment to BU-eLearning system (http://buelearning.hkbu.edu.hk/), under “COMPxxxx (1) [2022 S2] -> Group
Assignment” (where xxxx is the course code, i.e., 4135 or 7240).
o Submit the first two files under Turnitin assignment link “Submission of your project report and source files”: Part 1 for report (Word or PDF), and Part 2 for source files (in a .zip folder).
o Submit the video (no larger than 250MB in MP4 or MOV format) to another assignment link “Submission of your presentation video”.
o Note: if you meet any difficulties when you upload the assignment, please contact TA Ms. Wanling Cai cswlcai@Comp.HKBU.Edu.HK.
Assessment structure 系统开发代写
- This assignment covers 25% of the whole assessment.
- This is group work. It is expected that the workload is evenly shared by ALL the group members. The attached participation form should be filled in and signed by members.
- Assessment criteria: What we will be focusing on is not only the developed system itself, but also your abilities to design the user interface, implement the recommendation algorithms, and evaluate the results.
o Correctness: the recommendation algorithms are correctly reproduced and can work well in your delivered system.
o Completeness and interactivity: the required components are fully implemented and integrated. The whole system should look complete and interactive.
o Note: The implementations of core components should be different from those in our provided example prototype.
o Soundness of testing results: the system evaluation should be conducted in a rigorous procedure and all of the collected results should be analyzed in the right manner.
o Creativity: the novelties that you can show in the user interface design and algorithm implementation/improvement.
o Problem-solving: your abilities in solving the problems encountered during the system development and skills of self-learning additional references/materials.
o Presentation: your presentation skills in both written report and oral presentation.
o Team-working: individual members’ contributions to the whole project.
Cost of Plagiarism• Plagiarism is an offense and will result in appropriate disciplinary action against those involved.
- Penalty will be applied indiscriminatingly among those who involve (the one who copy and the one being copied). The minimum penalty would be receiving zero mark for the submitted work.
- Please refer to the following URL for the university’s guideline on penalizing plagiarism: http://ar.hkbu.edu.hk/curr/avoid_plagiarism/.
References 系统开发代写
- Datasets:
o http://eigentaste.berkeley.edu/dataset/
o https://guoguibing.github.io/librec/datasets.html
o https://grouplens.org/datasets/
o https://github.com/caserec/Datasets-for-Recommender-Systems
- Statistical methods:
o https://statisticsbyjim.com/hypothesis-testing/t-tests-excel/ (how to do tTests in Excel)
o https://www.reneshbedre.com/blog/ttest.html (how to do t-Tests in Python)
- Development toolkits:
o https://surprise.readthedocs.io/en/stable/index.html (Python library “Surprise” for the recommendation)
o https://vuejs.org/index.html (JavaScript Framework)
o https://github.com/axios/axios (Promise based HTTP client for the browser and node.js)
o https://element.eleme.cn/#/en-US (UI Library)
o https://fastapi.tiangolo.com/ (Web Framework)
