Description
One task is included in this assignment. The specification of the task starts in a separate page.
You must implement and run all your Python code in Jupyter Notebook. The deliverables include a project presentation, slides and source code.
All results of your implementation must be reproducible from your submitted Jupyter notebook source files. In addition, the submission must include all execution outputs as well as clear explanation of your implementation algorithms (e.g., in the Markdown format or as comments in your Python codes).
CSCI316 (SIM) 2022 Session 1 – Group Assignment 1
The Task
(10 marks)
Dataset: Steel Industry Energy Consumption Dataset
(Source: https://archive.ics.uci.edu/ml/datasets/Steel+Industry+Energy+Consumption+Dataset)
The information gathered is from the DAEWOO Steel Co. Ltd in Gwangyang, South Korea. It produces several types of coils, steel plates, and iron plates. The information on electricity consumption is held in a cloud-based system. The information on energy consumption of the industry is stored on the website of the Korea Electric Power Corporation (pccs.kepco.go.kr), and the perspectives on daily, monthly, and annual data are calculated and shown.
(Reference: Sathishkumar et al.’s Building Research & Information paper in the above link, which is accessible in the UOW digital library, i.e., https://www.uow.edu.au/library/)
Objective
The objective of this task is to develop an end-to-end data mining project by using the Python machine learning library Scikit-Learn. The output of the project is a classification model to predict the load type.
Requirements
- Main steps of the project are (a) “discover and visualise the data”, (b) “prepare the data for machine learning algorithms”, (c) “select and train models”, (d) “hyperparameter fine-tuning” and (e) “evaluate the outcomes”. You can structure the project in your own way. Some steps may be performed more than once.
- Clearly explain your findings at each step.
- In the steps (c) and (d), select and train at least 3 classifiers (from 3 different algorithms).
- Use ~80% data for training and ~20% for testing the models. Stratified sampling must be used.
- Define some new features by using the User-Defined Transform functionality, which implements a
parameter to use those new features or not in the model fine-tuning step (i.e., step (d)).
Deliverables
Deliverables include (1) a presentation of the project in an online tutorial and (2) a submission of the source code and slides via Moodle.
CSCI316 (SIM) 2022 Session 1 – Group Assignment 1
