Fundamentals of AI and ML
ML Problem Types
AIF-C01 expects you to match business problems to the right ML task. Focus on the output type: numeric value, class label, group, recommendation, unusual event, future time-series value, image/video insight, or language insight.
Concepts
- Regression predicts continuous numeric values such as price, demand, or claim amount.
- Classification predicts discrete classes such as fraud/not fraud, sentiment, or image label.
- Clustering discovers natural groups in unlabeled data, often for segmentation or pattern discovery.
- Recommendation predicts relevant items or content for a user based on behavior and item similarity.
- Anomaly detection identifies unusual observations that differ from expected patterns.
- Forecasting predicts future values in a time series using historical time-ordered data.
- Computer vision analyzes images or video; NLP analyzes, understands, or generates human language.
Exam tips
- Regression predicts numbers; classification predicts categories.
- Clustering is for unlabeled grouping, while recommendation ranks items for a user.
- Forecasting is time-series prediction; anomaly detection finds unusual records.