FIT5149 S2 2021 Assignment 2 – Marking Rubric Marking Rubric – Assignment 2 contributes 35% to your final unit mark Indicating the level the student is work at*: 0 column (N) Fail (P) Pass (C) Credit (D) Distinction (HD) High Distinction This assessment meets Unit Learning Outcomes 1,2,3,4,5 1. Analyze data sets with a range of statistical, graphical and machine-learning tools; 2. Evaluate the limitations, appropriateness, and benefits of data analytics methods for given tasks; 3. Design solutions to real-world problems with data analysis techniques; 4. Assess the results of the analysis; 5. Communicate the results of an analysis for both specific and broad audiences. Did not attempt Fail Pass Credit Distinction High Distinction Accuracy Did not attempt the assessment, or no better than random guess (0.5) Accuracy score is lower than 0.60 Accuracy score is larger than or equal 0.60 and lower than 0.72 Accuracy score is larger than or equal to 0.72 and lower than 0.77 Accuracy score is larger than or equal to 0.77 and lower than 0.80 Accuracy score is larger than or equal to 0.80 15 marks 0 0 – 7.5 7.5 – 9 9 – 10.5 10.5 – 12 12 – 15 Feature Extraction and Representation No attempt No or limited explanation of the features extraction process and text representation Sufficient explanation of the features extraction process, but the reasoning process of getting text representation is not well justified. A clear explanation of the features extraction process and the reasoning process of getting text representation is reasonably justified. The features extraction process is well explained and justified by the exact empirical result. The reasoning process of getting text representation is well justified. The features extraction process and reasoning process of getting text representation are well explained and justified by the exact empirical result. The explanation is easy to read for both machine learning expert and those who do not have the machine learning knowledge. 5 marks 0 0-2.5 2.5-3 3-3.5 3.5-4 4-5 Development of the classifier No attempt No, or limited explanation of how the classifier was developed with no evidence and proper justification. Limited explanation of how the classifier was developed with insufficient evidence and proper justification. Some explanation of how the classifier was developed with some relevant supporting information. Demonstrates some understanding of the difference between the compared models. A logical explanation of the classifier development process with supporting information. Demonstrates a solid understanding of the difference between the compared models. Critically assess the accuracy of the models with comprehensive analysis. The discussion of experiments and the comparison is logical and solid with an appropriate level of detail. The experimental methodology logically lead to the development of the final model to be submitted Demonstrates a high level of understanding of the performance differences between the candidate classifiers and those caused by different features. 10 marks 0 0-5 5-6 6-7 7-8 8-10 Code and Report Readability No commenting of code The R/Python code is incomplete and the code readability is poor. There is no clear structure in the report and the readability of the report is poor. The R/Python code is complete but poorly structured and the code readability is poor with no or insufficient comments to explain how the code is intended to work. The report is structured, but the The R/Python code is structured but the logic is not clear. Code readability is good with adequate comments to explain how the code is intended to work The report is well structured, and the readability of the report is reasonable. The R/Python code is logically structured and easy to read. Comments clearly explain how the code is intended to work. However, there are some redundant or unnecessary code and/or comments. The report is well structured and the The R/Python code is logically structured and easy to read. Concise but precise code comments clearly explain how the code is intended to work. The redundant or unnecessary code is excluded from the final submission. The README file is clear so that the submitted readability of the report is poor. readability of the report is good. implementation can be easily set up. The report is well structured and the readability of the report is good. 5 marks 0 0-2.5 2.5-3 3-3.5 3.5-4 4-5
