3.1 Probabilistic Modelling
Probabilistic Modelling: -
- Probabilistic modelling is a method in machine learning where probability is used to make predictions.
- Instead of giving a fixed answer, The model predicts the chance of something happening.
Example: Instead of saying “This email is spam”, the model says:
-
Spam → 90% probability
-
Not Spam → 10% probability
So the system decides that the email is most likely spam.
- Real-world data always contains uncertainty and noise. so Probabilistic models help machines handle uncertainty and make better decisions.
They are used for:
-
Email spam detection
-
Weather prediction
-
Medical diagnosis
-
Fraud detection
-
Recommendation systems
Probabilistic Models Working steps
1. Collect Data
Example: past weather data.
2. Find Probability Patterns
The model learns how often events happen.
3. Build a Mathematical Model
It calculates probabilities.
4. Make Predictions
It predicts the most probable outcome.
Simple Probability Formula
P(A) = {Number of favourable outcomes}/{Total number of outcomes}
Where:
-
P(A) = Probability of event A
-
Example: probability of getting a head in a coin toss
P(Head) = 1 / 2 = 0.5
Common Probabilistic Models in Machine Learning
Some popular probabilistic algorithms:
-
Naive Bayes
-
Used for spam detection
-
Works using probability rules
-
-
Bayesian Networks
-
Used for medical diagnosis
-
-
Hidden Markov Models
-
Used in speech recognition
-
-
Gaussian Mixture Models
-
Used in clustering problems
-