Reference no: EM133745283 , Length: word count:1000

Artificial Intelligence and Machine Learning

Assessment - Case Study and Individual Exercise

Your Task
Group: Using the Orange Data Mining software, implement prediction workflows based on decision tree and logistic regression models to perform a machine learning task. Answer the accompanying questions on decision tree and logistic regression models. You can get help from your group members, but you must implement your own prediction workflow and answer the accompanying questions yourself.

Assessment Description

Background: We have already explored several machine learning models, such as decision trees and logistic regression. It is possible to implement prediction workflows based on these models using a tool such as Orange, which has a widget-driven interface.
In this assessment, you will create machine learning models that can make predictions based on the data and report on the characteristics of the models and their potential for prediction.

Assessment Instructions

Implementation (in class). As a group,
Use Orange to implement prediction workflows based on decision trees and logistic regression to perform machine learning tasks. The dataset and parameters are given in the assessment sheet.

Based on the Orange output, answer the accompanying questions in the assessment sheet on the prediction accuracies of the decision tree and logistic regression models.

Report (after class). As an individual, take notes in class and then write a 1000-word (maximum) report that summarises and provides suggestions for further analysis. This summary report is a part of the assessment and must be submitted via Turnitin. Additionally, the Orange workflow file must be submitted to the file submission box.

Group Activity (in-class)

Implement a prediction model using Orange.

Solve problems based on the prediction model.

Present the work to demonstrate understanding.

Specific Details:

A predictive model has been built to the quality where it can be used for analysis. This should include:

Data loaded into Orange Data Mining and processed.
Appropriate ML models developed to perform a classification task.

An analysis of the data has been made manually to understand and verify the output of the Orange software:

Questions on the accompanying assessment sheet answered.
Using collaboration, via discussion to complete the task (as observed by the assessor).

The analysis findings have been verbally presented to the facilitator. A successful findings report should include:

All members of the group reporting a part of the findings,
An explanation of methodology provided

Individual Report and Summary

Data characteristics and quality summarised.

Plots/charts described and justified.

Feasible recommendations for improving predictive method provided.

Well-structured and readable report with relevant headings. The report provides an organisational use-case.

Justification of assumptions and interpretations are clear and concise.

In-line referencing used and references are relevant and genuine.

Visualisations are used to convey key arguments.

Specific Details:

Succinct language has chosen to adhere to a word limit without diluting meaning.

An accurate summary of the presentation information has been provided that would help someone who did not see the oral presentation have a clear understanding of the key findings and insights.

At least 2 additional techniques predictive have been suggested, indicating a solid grasp of standard predictive models.

An informal report format has been used which contains readability standards that would be appropriate for business use between colleagues. A successful report could consist of:

Language clarity and accuracy
Headings for easy identification and readability
Bullet points for succinctness and readability
Punctuation for expression and readability
Referencing for authenticity and appropriate source recognition