🚀 AI Terminologies Every Beginner Should Know (Part 2)

If you're learning AI, these are some of the most common terms you'll encounter. Understanding them early will make advanced topics much easier.

1. Dataset
A collection of data used to train, validate, or test an AI model.
Example: A folder containing 50,000 images of cats and dogs.

2. Training Data
The data used to teach an AI model how to perform a task.
Example: Thousands of emails labeled as "Spam" or "Not Spam."

3. Test Data
New, unseen data used to evaluate how well a trained model performs.

4. Features
The input variables or characteristics used by an AI model to make predictions.
Example: Age, salary, and years of experience for predicting employee attrition.

5. Labels
The correct answers or target values that the model learns to predict.
Example: "Approved" or "Rejected" in a loan prediction dataset.

6. Model
A trained AI system that has learned patterns from data and can make predictions or generate outputs.

7. Algorithm
A set of rules or mathematical procedures used to train an AI model.
Examples: Linear Regression, Decision Tree, Random Forest.

8. Parameters
The values learned by a model during training.
These determine how the model makes predictions.

9. Hyperparameters
Settings chosen before training begins.
Examples:
• Learning Rate
• Batch Size
• Number of Epochs

10. Epoch
One complete pass of the entire training dataset through the model.
If you train for 20 epochs, the model has seen the complete dataset 20 times.

11. Batch
A small subset of training data processed at one time.
Instead of training on 100,000 records together, the model may process batches of 32 or 64 records.

12. Loss Function
A mathematical function that measures how wrong the model's predictions are.
Lower loss generally means better performance.

13. Optimization
The process of updating model parameters to reduce the loss.

14. Learning Rate
Controls how big each update is while training the model.
• Too high → Model may overshoot.
• Too low → Training becomes very slow.

15. Accuracy
The percentage of correct predictions made by a model.
Example:
If a model correctly predicts 95 out of 100 cases, its accuracy is 95%.

16. Precision
Out of all positive predictions, how many were actually correct.

17. Recall
Out of all actual positive cases, how many the model correctly identified.

18. F1 Score
A balanced metric that combines Precision and Recall into a single score.

19. Confusion Matrix
A table used to evaluate classification models by showing:
• True Positives
• False Positives
• True Negatives
• False Negatives

20. Prediction
The final output generated by an AI model after processing new data.

Example:
Predicting whether a customer will churn or whether an email is spam.

❤️ Double tap for more