Summary Cyber Security notes - BCA with AI,ML and data science

LESSON [5-2]: AI - ML - EVALUATING MODELS
A model that performs
Good Model
reliably for its intended task.
Assessment of a model's
Model Quality
effectiveness and reliability.

Classification Model
A model used to categorize
data into classes.

Classifier Performance
Evaluation of how well a
classifier pre-dicts outcomes.

Misclassification Error
Incorrectly classifying a
record into the wrong class.
Percentage of misclassified
Error Rate
records in validation data.
Probability of correct predictions
Accuracy by the model; looks at overall
perfor-mance (1 - Error Rate)

Table summarizing binary classifica-
Confusion Matrix
tion errors.

False Positive
1/7

, LESSON [5-2]: AI - ML - EVALUATING MODELS
Incorrectly predicted positive event

occurrence.
Incorrectly predicted
False Negative
non-event oc-currence.
- Proportion of actual
True Positive Rate
positives cor-rectly predicted.
Of all times the event occurred,
TPR Hit Rate what is the proportion of time it
was predict-ed to occur?

Model's ability to
TPR Sensitivity
identify positive events.
Proportion of actual negatives
True Negative Rate (TNR)
correct-ly predicted.
- The proportion of non-
events cor-rectly predicted.
TNR Specificity - Of all the times the event did not

oc-cur, what was the proportion of

times it was predicted to not occur?


2/7