Business Analytics 101: Machine Learning Fundamentals Supervised vs Unsupervised Learning

What This Is

Supervised and unsupervised learning are two fundamental types of machine learning algorithms used in business analytics. Supervised learning involves training a model on labeled data to make predictions on new, unseen data. Unsupervised learning, on the other hand, involves discovering hidden patterns or relationships in unlabeled data. Both types of learning are crucial in business analytics, as they enable organizations to make informed decisions and drive business growth. For instance, a retail company can use supervised learning to predict customer churn based on historical data, while using unsupervised learning to segment customers based on their purchasing behavior.

Key Formulas & Metrics

• Mean Absolute Error (MAE) = (1/n) Σ|yᵢ − ŷᵢ|: average absolute forecast error, where yᵢ is the actual value, ŷᵢ is the predicted value, and n is the number of observations.
• R² = 1 − (SS_res / SS_tot): proportion of variance explained by the regression model, where SS_res is the sum of squared residuals and SS_tot is the total sum of squares.
• Coefficient of Determination (R²) = 1 - (Σ(yᵢ - ŷᵢ)² / Σ(yᵢ - ȳ)²): measures the goodness of fit of a regression model, where yᵢ is the actual value, ŷᵢ is the predicted value, and ȳ is the mean of the actual values.
• Mean Squared Error (MSE) = (1/n) Σ(yᵢ - ŷᵢ)²: average squared difference between actual and predicted values.
• Root Mean Squared Error (RMSE) = √(MSE): square root of the average squared difference between actual and predicted values.
• Akaike Information Criterion (AIC) = n * ln(MSE) + 2 * k: measures the relative quality of a model, where n is the number of observations, MSE is the mean squared error, and k is the number of parameters.
• Bayesian Information Criterion (BIC) = n * ln(MSE) + k * ln(n): measures the relative quality of a model, where n is the number of observations, MSE is the mean squared error, and k is the number of parameters.

Step-by-Step Procedure

1. Define the problem: Clearly articulate the business problem you're trying to solve, such as predicting customer churn or segmenting customers.
2. Collect and preprocess data: Gather relevant data, handle missing values, and transform variables as necessary.
3. Split data into training and testing sets: Divide data into training and testing sets to evaluate model performance.
4. Choose a model: Select a supervised or unsupervised learning algorithm based on the problem and data.
5. Train and evaluate the model: Train the model on the training data and evaluate its performance on the testing data.
6. Interpret results: Analyze the model's performance metrics, such as MAE, R², and AIC, to determine its effectiveness.

Common Mistakes

• Mistake: Confusing correlation with causation.
• Correction: Correlation does not imply causation. Establish a clear cause-and-effect relationship between variables.
• Mistake: Misinterpreting p-values.
• Correction: A small p-value does not necessarily mean that the null hypothesis is false. Consider the context and other evidence.
• Mistake: Using the wrong error metric for a business problem.
• Correction: Choose an error metric that aligns with the business objective, such as MAE for forecasting or R² for regression.

Software / Tool Tips

• Python with scikit-learn: Use the LinearRegression class for linear regression and the KMeans class for k-means clustering.
• R: Use the lm() function for linear regression and the kmeans() function for k-means clustering.
• Tableau: Use the Regression and Clustering tools to create visualizations and perform analysis.

Quick Practice Problem

Problem: Compute the moving average forecast for period 4 given data: 100, 110, 105, 115.

Answer: The moving average forecast for period 4 is (100 + 110 + 105 + 115) / 4 = 110.

Explanation: The moving average forecast is the average of the last four periods.

Last-Minute Cram Sheet

1. Supervised learning involves training a model on labeled data to make predictions on new, unseen data.
2. Unsupervised learning involves discovering hidden patterns or relationships in unlabeled data.
3. MAE measures the average absolute difference between actual and predicted values.
4. R² measures the proportion of variance explained by a regression model.
5. AIC measures the relative quality of a model.
6. BIC measures the relative quality of a model.
7. Correlation does not imply causation.
8. A small p-value does not necessarily mean that the null hypothesis is false.
9. Choose an error metric that aligns with the business objective.
10. ⚠️ p-value is NOT the probability that H₀ is true – it’s the probability of observing the data (or more extreme) if H₀ is true.