Agent is an entity that can perceive its environment through sensors

and act upon that environment using actuators. Agents are designed
to achieve specific goals or perform tasks based on their observations
and the knowledge they acquire. They are fundamental to AI systems
and are used to model decision-making processes, automation,
and problem-solving.
Types-Reactive Agents,Model-Based Agents,Goal-Based Agents,
Utility-Based Agents,Learning Agents. Components-Perception,
Environment,Actuators,Decision-Making Module.
Support Vector Machine (SVM) is a powerful supervised learning
algorithm used for classification and regression tasks. It is widely
favored due to its effectiveness in handling high-dimensional data
and its robustness against overfitting, especially in cases where
the number of dimensions is greater than the number of samples.
Key Concepts-Hyperplane,Margin,Support Vectors,Advantages of SVM
Effective in high-dimensional spaces. Works well with a clear margin of
separation. Memory-efficient as it only uses support vectors.
Can handle non-linear relationships using kernels//Not ideal for large
datasets to high computational cost. Applications -Text Classification,
Image Recognition,Fraud Detection
Principal Component Analysis (PCA) is an unsupervised
machine learning technique used for dimensionality reduction.
It transforms a high-dimensional dataset into a smaller one by
finding new axes (principal components) that maximize the
variance of the data while retaining most of the important
information. Key Concepts of PCA-Dimensionality Reduction::-
Reduces the number of features in a dataset by transforming
the data into a lower-dimensional space.
Principal Components--Linear combinations of the original
features, ordered by the amount of variance they capture.
Steps-1.Standardize the Data2.Compute the Covariance
Matrix 3.Compute Eigenvectors and Eigenvalues 4.Sort and
Select Principal Components 5.Transform the Data....Advantages of PCA
Reduces overfitting by eliminating redundant features.
Improves computational efficiency for high-dimensional data.
Helps visualize data in lower dimensions (2D or 3D).
Diasadvantage-Loses interpretability of original features.
Assumes linear relationships between variables.
Gradient Descent is an optimization algorithm used in machine
learning to minimize a function by iteratively moving in the direction
of the steepest descent, as defined by the negative gradient. It is
widely used to optimize loss functions in supervised learning
algorithms like linear regression, logistic regression, and neural networks.
Key Concepts-Objective Function,Gradient-A vector of partial
derivatives of the function with respect to its parameters.,Learning
Rate (α). 1.Type-Batch Gradient DescentComputes the gradient using
the entire dataset.Pros: Accurate updates.Cons: Slow for large datasets.
2.Stochastic Gradient Descent (SGD) 3.Mini-Batch Gradient Descent
Stochastic Gradient Descent (SGD) is a variant of the gradient descent
optimization algorithm. Instead of computing the gradient of the loss
function for the entire dataset (as in Batch Gradient Descent), SGD
updates the model parameters for each training sample individually.
This makes it faster for large datasets and suitable for online learning.
Random Forest is a popular ensemble machine learning algorithm
that combines multiple decision trees to improve prediction accuracy
and reduce overfitting. It is used for both classification and
regression tasks. The algorithm builds a "forest" of decision trees
and aggregates their predictions for final results.
ROC-When dealing with imbalanced datasets in classification
problems, Random Forest can be enhanced by evaluating model
performance using metrics like ROC (Receiver Operating
Characteristic), Precision, and Recall.When dealing with
imbalanced datasets, metrics like ROC-AUC, Precision,
and Recall provide a better understanding of model performance
than accuracy. Forward Chaining is a data-driven inference
method where the reasoning starts with the available facts and
applies rules to infer new facts until the goal is reached.
Process:-The system begins with known facts. It iteratively
applies rules to these facts to infer new facts.The process
continues until a goal or desired conclusion is achieved.
Backward Chaining-is a goal-driven inference method where
the reasoning starts with the desired goal and works backward
to see if there is evidence to support that goal.
Process:-The system starts with a goal (or hypothesis).
It looks for rules that would support this goal.
Partial Order Planning (POP) is a form of automated planning
used in Artificial Intelligence to find a sequence of actions that
achieve a goal, but unlike traditional total order planning,
it doesn't impose a strict linear sequence on the actions. Instead,
it maintains a partial order where only the necessary dependencies
between actions are explicitly ordered, allowing greater flexibility in
the plan. Steps-1.Start with the Initial State 2.Identify the Goal,
3.Generate Actions, 4.Plan Optimization
A search algorithm* is a widely used pathfinding and graph
traversal algorithm in AI. It is designed to efficiently find the
shortest path from a starting node to a goal node in a weighted
graph, making it ideal for applications such as robotics, game
AI, and navigation systems.
Lasso Regression (L1 Regularization)
Objective: Similar to Ridge, but the penalty term is the sum of the
absolute values of the coefficients.Ridge Regression (L2 Regularization)
Objective: Minimize the sum of squared residuals, plus a penalty term that
is proportional to the square of the magnitude of coefficients.Ridge is
useful when you want to keep all features but control their impact.
Lasso is helpful when you want to identify a subset of features that
are the most important.
activation function in a neural network determines whether
a neuron should be activated or not by calculating the
weighted sum of inputs and applying a non-linear transformation.
It introduces non-linearity, allowing neural networks to
learn complex patterns.
Sigmoid and Tanh are best for smaller networks and binary
classification tasks. ReLU is generally preferred for deeper networks
and hidden layers due to its efficiency. Softmax is used in multi-class
classification to output probabilities. Leaky ReLU or Swish can be
used when ReLU causes issues like dying neurons.