Can't post file, will upload on next page
ASSESSMENT GUIDE Unit Code: ITEC203 Introduction to Data Science and Machine Learning Study Period: S1 2022 Assessment number (3) Assessment Artefact: Code and Presentation Weighting [40%] Why this assessment? What are the types of employability skills that I will acquire upon completion of this assessment? Assessment Overview: Purpose, as written in the EUO Due date: 17/06/2022, 5pm, on Friday of Week 15 Weighting: 40% • Opportunity to apply theory into practice • Exposure to real-life scenario • Enhance the understanding of theoretical concepts introduced in workshops – the definition of each problem, the logic of the algorithms, the meaning of their parameters and hyperparameters, the cons and pros of them. • Build up students’ coding and problem-solving skills with coding practice. • Build up students’ documentation skills with report output. • The feedback from this assessment will help students to be ready to correct any conceptual misunderstanding and apply in more scenarios when taking roles e.g. data engineer in the job markets. Skill Type Developed critical and analytical thinking ☑ Developed ability to solve complex problems ☑ Developed ability to work effectively with others ☑ Developed confidence to learn independently ☑ Developed written communication skills ☑ Developed spoken communication skills ☑ Developed knowledge in the field study ☑ Developed work-related knowledge and skills ☑ 2 Length and/or format: Document in the format of Jupyter notebook including Python source codes and recorded presentation – about 5 minutes Learning outcomes assessed LO3, LO4 Graduate attributes assessed GA2, GA5, GA8 How to submit: Via LEO Return of assignment: Via LEO as final marks Assessment criteria: Rubric: see end of document Context Suppose the students take a Data Engineer role in a company. Doing supervised learning, e.g., classification, would be a daily task for some of the real-world projects. So this assignment would enhance the students’ understanding on the theoretical knowledge introduced in the workshops. In addition, from those hands-on practice, students would gain deeper insights on how to apply machine learning algorithms to solve real-world problems. Instructions Provide instructions for students for completing assignment This assignment will focus on applying two Multiclass classification algorithms on the MNIST datasets. Tasks of this assignment include 1. Split the whole dataset into training/testing dataset either by yourself or by functions provided by python scikit-learn packages, explain in details what your code is doing. (5%) 2. Exploit the performance of kNN classifier on the dataset (15%) a. Set k=10 which is the ground-truth number of categories in the dataset, then apply kNN classifier (using functions from Python package), explain in detail the meaning of parameters for the function interface you choose. (5%) b. Evaluate the performance with the metrics introduced in workshop (also using the metric functions provided by Python scikit-learn package), explain in detail why a specific evaluation metric is picked and what you can tell from the results about the model. (5%) c. Experimenting with different values of k, and comparing the performance. What can you discover from the comparison? Is there any inspiration on choosing k? (5%) 3. Exploit the performance of SVM classifier on the dataset (10%) a. Try appropriate SVM classifiers from https://scikit-learn.org/stable/modules/svm.html, and explain in details the meaning of parameters of the function interface you choose. (5%) b. Compare the performance of the SVM classifier and the kNN classifier. What have you discovered from such comparison, such as any clues for your future decision on what model to use in what situation? (5%) 4. record presentation: using succinct language to explain what you’ve done and what you’ve discovered in 5 minutes. (10%) Structure Prepare a Jupyter notebook and video recordings for this assignment. The structure of the Jupyter notebook should alternate texts and python codes and cover topics listed the in specific tasks above. The video recording is used to confirm academic integrity. Each cell in the Jupyter notebook needs to be explained in the video. How do I submit? Submit to Assessment 3 via LEO assessment tile Note that: To make sure the submission satisfies academic integrity, the code will be compared to other students’ https://scikit-learn.org/stable/modules/svm.html 3 submission in Turnitin. Submission checklist I have formatted my report as per the specifications ☐ I have checked my Turnitin report and taken appropriate actions to ensure that the submission satisfies academic integrity ☐ I have actioned feedback advice provided to me from labs feedback and assignment 2 (if applicable) ☐ I have submitted my work before the due date/time ☐ I have submitted feed forward template along with my assignment submission ☐ Feed Forward Template (example) A template for students to use and act on feedback and provide recommendations for improvement. Note This is a task for any instance of follow-on assignment (assessment 2 and 3). This must be submitted as the first page of the follow-on assignment (assessment 2 and 3) to ensure you acted on the feedback provided to you in the previous assignment (this is not counted as part of the assessment word count). How did you act on the feedback? Feedback is an important component of learning. Please consider the feedback you received in your last assignment and provide a response on how you acted on, or intend to act upon, that feedback, and how it has informed the current assignment task. Submit this sheet along with your assignment. Questions Your learning from the previous assignment feedback How have you acted on the feedback from previous assignment to improve your work in this assignment? (e.g. based on my previous feedback, I made sure that I supported my discussion, position, ideas, concepts with peer reviewed journal references in this assignment) What is your expectation around the type of feedback that enhances your learning? (e.g. I want to know where I made a mistake and how I can correct them and not make the same mistake again i.e. I want specific feedback that will help me to improve my learning and performance in the next assignment) Did you have any difficulty understanding or acting on previous feedback? Please be as specific as possible so that you can gain further feedback/clarify anything you do not understand in the feedback (e.g. feedback provided in my previous assignment was very generic I did not know how to improve my work. So, I would like the teacher to explain more on xxxx aspects of the feedback or I would like an opportunity to have a dialogue to understand the feedback) 4 Some Helpful Websites and Resources Add in a couple of places to go for more info Model Evaluation: https://scikit-learn.org/stable/auto_examples/model_selection/plot_roc.html Cross-validation: https://scikit-learn.org/stable/modules/cross_validation.html Grid search: https://scikit-learn.org/stable/modules/grid_search.html Who can help me? Studiosity Academic skills Unit (ASU) Places – NLiC Maoying Qiao (
[email protected]) LiC Zijing Chen (
[email protected]) Online Facilitator Maoying Qiao Lab demonstrator Zijing Chen I’m having problems Special Consideration: This form is used by students to apply for Special Consideration for assessable work in studies at Australian Catholic University. Approval of such applications will only be granted to students who are legitimately disadvantaged in their assessment due to exceptional and unforeseen circumstances beyond their control. Referencing All referencing should be in ACU Harvard style; however if you are coming from another faculty, you may choose to use your usual referencing style. If this is the case you must indicate at the top of your reference list what referencing style you are using (e.g. APA, MLA, Chicago, etc). Please ensure your assignment makes use of in-text citations and a reference list. Missing citations or references is equivalent to plagiarism. Criteria The full criteria is compiled in a rubric, which can be found on the following page/s. https://scikit-learn.org/stable/auto_examples/model_selection/plot_roc.html mailto:
[email protected] https://units.acu.edu.au/__data/assets/word_doc/0006/620655/SC_Application_for_Special_Consideration_20180214.docx https://libguides.acu.edu.au/referencing/harvard 5 Rubric for [insert assessment item title and weighting] Relevant LO/GAs Criterion (related to a single GA from the related LO – one GA per criterion Does not meet expectations Meets expectations Exceeds expectations NN PA CR DI HD GA5 LO3 Weight=15 marks TL=3 Learning stage = I and D Demonstrate correct understanding of the concepts of classification, kNN, SVM, training-testing splitting, evaluation Fail to adequately demonstrate correct understanding of the concepts of classification, kNN, SVM, training-testing splitting, evaluation (0 – 7.35) Adequately demonstrate correct understanding of the concepts of classification, kNN, SVM, training-testing splitting, evaluation (7.5 – 9.6) Credibly demonstrate correct understanding of the concepts of classification, kNN, SVM, training-testing splitting, evaluation (9.75 – 11.1) Distinctively demonstrate correct understanding of the concepts of classification, kNN, SVM, training-testing splitting, evaluation (11.25 – 12.6) Highly distinctively demonstrate correct understanding of the concepts of classification, kNN, SVM, training-testing splitting, evaluation (12.75 – 15) GA8 LO3 Weight=15 marks TL=3 Learning stage = I and D Demonstrate python programming skills by experimenting with related functions in scikit-learn package for classification and evaluation as well as model selection. Fail to adequately demonstrate python programming skills by experimenting with related functions in scikit-learn package for classification and evaluation as well as model selection. (0 – 7.35) Adequately demonstrate python programming skills by experimenting with related functions in scikit-learn package for classification and evaluation as well as model selection. (7.5 – 9.6) Credibly demonstrate python programming skills by experimenting with related functions in scikit-learn package for classification and evaluation as well as model selection. (9.75 – 11.1) Distinctively demonstrate python programming skills by experimenting with related functions in scikit-learn package for classification and evaluation as well as model selection. (11.25 – 12.6) Highly distinctively demonstrate