Machine Learning Algorithms: Machine Learning Algorithm Evaluation

This post covers Machine Learning Algorithms, as detailed below.

BBIM708 Machine Learning and AI Assessment 2: Algorithm Evaluation and Fine-Tuning Portfolio

This NZSE portfolio is a hands-on machine learning task. In Part A you pick one supervised and one unsupervised algorithm, apply them to an AI problem of your choice with open-source datasets, and evaluate them with a diagnostic visualisation and two metrics. In Part B you optimise a classification model and a regression model using Grid Search, Random Search and 5-fold cross-validation, with screenshots of all code and results.

Part A: Evaluating ML Algorithms (44 marks, LO3)

Select one supervised algorithm (classification or regression) and one unsupervised algorithm. Apply each to a specific AI application (for example house price prediction, customer segmentation or fraud detection) using relevant open-source datasets, then train, evaluate and analyse the models.

1. Working principles (4 marks): explain each algorithm’s key computational steps and theoretical foundations, focusing on concepts rather than code (about 100 words per algorithm).
2. Suitable AI problems and data (4 marks): identify and justify the types of AI problems and data each algorithm suits best (about 100 words per algorithm).
3. Strengths and limitations (4 marks): discuss two strengths and two limitations of each algorithm with practical examples (about 150 words per algorithm).
4. Train the models (4 marks): screenshots of the training code with brief explanations.
5. Diagnostic visualisation (10 marks): one visualisation per model, such as a confusion matrix, ROC curve or cluster plot, with a 75-word interpretation and code screenshots.
6. Performance metrics (10 marks): evaluate each model with two appropriate metrics (accuracy, precision, F1-score, RMSE and so on) shown in tables or charts with code screenshots.
7. Interpret the results (8 marks): about 150 words per model on performance and practical relevance.

Part B: Optimising ML Models with Fine-Tuning Techniques (56 marks, LO4)

Choose supervised classification and regression problems using datasets from sklearn, Kaggle, UCI or your own source, and optimise model performance using Grid Search, Random Search and cross-validation.

1. Grid Search on a classification model (18 marks): choose a model such as Logistic Regression, Random Forest or SVM, select two hyperparameters with two values each, test every combination and report the best one with Accuracy, Precision, Recall or F1-score. Evidence: model and dataset justification (150 words), annotated Grid Search code (150 words), a hyperparameter combinations table, a best-configuration summary and a performance comparison discussion.
2. Random Search on a regression model (18 marks): choose a model such as Linear Regression, Decision Tree Regressor or Random Forest Regressor, define broader ranges for two hyperparameters, sample random combinations and report the best one using RMSE, R² or MAE, with the same evidence set as Task a.
3. K-Fold Cross-Validation (20 marks): apply 5-fold cross-validation to both models, reporting Accuracy and F1-score per fold for classification and RMSE and R² per fold for regression, plus the average across folds. Evidence: annotated code (150 words), a results table for all five folds and a 200-word interpretation of consistency, generalisability and reliability.

Submission Notes

• Submit a Word document or PDF with screenshots of all programming scripts and their outputs, plus compressed files of the scripts and data, on Canvas.
• Answers must be in your own words; AI writing and rephrasing tools are not permitted, and Turnitin checks similarity and AI content.
• No more than 15% of the content may be quoted or paraphrased, with APA 7 citations. Use Calibri or Times New Roman 12pt, 1.5 line spacing and page numbers.

The post BBIM708 Machine Learning and AI Assessment 2 Portfolio 2026: Supervised and Unsupervised Algorithms, Grid Search, Random Search and K-Fold Cross-Validation | NZSE appeared first on NZ Assignment Help.

For further reading, see this external resource.

You can also explore more resources on our website.