Machine Learning Course

Section 4 — Introduction to Optimisation and the Gradient Descent Algorithm

• Cost Functions
• Creating Charts with Subplots
• Gradient Descent Algorithm
• Optimisation
• Understanding the Learning Rate
• Creating 3-Dimensional Charts
• Understanding Partial Derivatives and How to use SymPy
• Batch Gradient Descent with SymPy
• Concatenating Numpy Arrays
• The Mean Squared Error (MSE)
• A MSE Cost Function
• Gradient Descent with a MSE Cost Function

Section 5 — Predict House Prices with Multivariable Linear Regression

• Index Data, Pandas Series, and Dummy Variables
• Descriptive Statistics: the Mean vs the Median
• Correlation: Understanding Strength & Direction
• Calculating Correlations and the Problem posed by Multicollinearity
• Visualizing Correlations with a Heatmap
• Seaborn Pairplots & Jupyter Microbenchmarking Techniques
• Multivariable Regression
• How to Shuffle and Split Training & Testing Data
• How to Calculate the Model Fit with R-Squared
• Model Evaluation
• How to Interpret Coefficients using p-Values and Statistical Significance
• VIF & Testing for Multicollinearity
• Model Simplification & Bayesian Information Criterion
• How to Analyze and Plot Regression Residuals
• Build a Valuation Tool: Working with Pandas Series & Numpy ndarrays

Sections 6 — Pre-Process Text Data for a Naive Bayes Classifier to Filter Spam Emails: Part 1

• The Naive Bayes Algorithm and the Decision Boundary for a Classifier
• Basic Probability
• Joint & Conditional Probability
• Bayes Theorem
• Introduction to Natural Language Processing (NLP)
• Tokenizing, Removing Stop Words and the Python Set Data Structure
• Word Stemming & Removing Punctuation
• Removing HTML tags with BeautifulSoup
• Advanced Subsetting on DataFrames: the apply() Function
• Logical Operators to Create Subsets and Indices
• Word Cloud with Python
• Sparse Matrix (Part 1): Split the Training and Testing Data
• Sparse Matrix (Part 2): Data Munging with Nested Loops
• Sparse Matrix (Part 3): Using groupby() and Saving .txt Files

Sections 7 — Train a Naive Bayes Classifier to Create a Spam Filter: Part 2

• Create a Full Matrix
• Count the Tokens to Train the Naive Bayes Model

Sections 8 — Test and Evaluate a Naive Bayes Classifier: Part 3

• Joint Conditional Probability (Part 1): Dot Product
• Joint Conditional Probability (Part 2): Priors
• Making Predictions: Comparing Joint Probabilities
• The Accuracy Metric
• False Positive vs False Negatives
• The Recall Metric, Precision Metric, F-score or F1 Metric
• A Naive Bayes Implementation using SciKit Learn