0. ML Overview/0.2.0 ML Intro - Categories of ML Algos.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "9ca8c6c4-a7d3-4ac2-8f0c-1f10c2ffa229",
   "metadata": {},
   "source": [
    "# Categorization of Machine Learning Algorithms\n",
    "\n",
    "Machine learning algorithms can be categorized in various ways, each focusing on different aspects. Understanding these categories is crucial for selecting the appropriate algorithm for a given task. Below are some of the key ways to categorize machine learning algorithms:\n",
    "\n",
    "#### 1. Based on Learning Paradigm\n",
    "\n",
    "- **Supervised Learning**: Algorithms that learn from labeled training data to make predictions.\n",
    "- **Unsupervised Learning**: Algorithms that infer patterns from unlabeled data.\n",
    "- **Semi-Supervised Learning**: A mix of supervised and unsupervised learning methods.\n",
    "- **Reinforcement Learning**: Algorithms that learn by taking actions and receiving feedback from the environment.\n",
    "\n",
    "#### 2. Based on an Algorithmic Approach or Model\n",
    "\n",
    "- **Linear and Non-linear Regression**: Methods for modeling relationships in data, ranging from simple linear to complex non-linear relationships.\n",
    "- **Regularized and Bayesian Regression**: Includes techniques for managing overfitting and incorporating probabilistic approaches in regression analysis.\n",
    "- **Deep Learning**: Involves algorithms based on artificial neural networks with multiple layers.\n",
    "- **Decision Trees**: Tree-like models of decisions and their consequences.\n",
    "- **Probabilistic Methods**: Algorithms using statistical approaches for modeling and inference.\n",
    "- **Ensemble Methods**: Techniques that combine multiple machine learning algorithms to improve performance.\n",
    "\n",
    "#### 3. Based on Functionality or Taskn\r\n",
    "\n",
    "\n",
    "- **Classification**: Assigning data points to predefined categories.\n",
    "- **Regression**: Predicting continuous values.\n",
    "- **Clustering**: Grouping similar data points together.\n",
    "- **Association**: Discovering rules that capture relationships between variables in data.\n",
    "- **Dimensionality Reduction**: Reducing the number of features in data while retaining important information.\n",
    "- **Generation**: Create new information based on the input of the user.\n",
    "\n",
    "#### 4. Based on Learning Technique\n",
    "\n",
    "- **Batch Learning**: The model is trained using the entire dataset at once.\n",
    "- **Online Learning**: The model is trained incrementally as new data arrives.\n",
    "\n",
    "#### 5. Based on the Type of Data Handled\n",
    "\n",
    "- **Time-Series Analysis**: For time-dependent data.\n",
    "- **Text Analysis and NLP**: Focused on processing and analyzing text data.\n",
    "- **Graph Algorithms**: For data modeled in graph structures.\n",
    "\n",
    "#### 6. Additional Categorizations\n",
    "\n",
    "- **Evolutionary Algorithms**: Based on natural selection concepts.\n",
    "- **Fuzzy Systems**: Utilizing fuzzy logic for approximate reasoning.\n",
    "- **Spectral Methods**: Involving matrix eigenvalues and eigenvectors.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a72f4eab-2f03-4b33-88d0-eca993a62389",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d822bf95-6157-4f97-892e-2170797e4ee3",
   "metadata": {},
   "source": [
    "### 1. Based on Learning Paradigm\n",
    "\n",
    "\n",
    "#### What is Supervised Learning?\n",
    "Supervised Machine Learning involves training a model on a labeled dataset, where the data includes both the input (features) and the correct output (labels). The goal is for the model to learn to predict the output based on the input data.\n",
    "- **Basic Concept**: The algorithm learns from data that includes both features and labels, like a student learning with the help of a teacher who provides homework (data) and solutions (labels).\n",
    "- **What It Is**: This approach is akin to learning under guidance, where the 'teacher' (data scientist) ensures the machine is trained on unbiased data, encompassing both questions (features) and answers (labels), to enable accurate predictions.\n",
    "- **Examples and Use Cases**: Common applications include predicting house prices using known price data and features like size and location, or classifying emails as spam or non-spam. It's suitable for scenarios with clear, labeled data for predicting outcomes on new, similar datasets.\n",
    "- **Training Process**: The learning algorithm analyzes the training data and develops a function to map inputs to outputs. The goal is to make accurate predictions or decisions.\n",
    "- **Feedback Loop**: It involves a feedback mechanism where the algorithm's predictions are compared against the actual outcomes to find errors and make adjustments.\n",
    "\n",
    "\n",
    "#### What is Unsupervised Learning?\n",
    "Unsupervised Learning involves algorithms that infer patterns from unlabeled data, without explicit instructions on what to do with it.\n",
    "- **Basic Concept**: In Unsupervised Learning, the algorithm is given data without any explicit instructions on what to do with it. The data does not have labels or annotations.\n",
    "- **What It Is**: Unsupervised learning is like learning without a teacher. The machine tries to find patterns and relationships in data by itself.\n",
    "- **Examples and Use Cases**: Grouping customers into segments based on purchasing behavior, or finding associations in shopping patterns.\n",
    "- **Pattern Discovery**: The main goal is to discover patterns and relationships in the data. The algorithm tries to organize the data in some way or describe its structure.\n",
    "- **Approaches**: Common approaches include clustering (grouping similar items) and association (discovering rules that describe parts of the data).\n",
    "\n",
    "\n",
    "#### What is Semi-Supervised Learning?\n",
    "Semi-Supervised Learning is a combination of supervised and unsupervised learning methods. It involves using both labeled and unlabeled data for training.\n",
    "- **Basic Concept**: Semi-Supervised Learning uses both labeled and unlabeled data for training, typically with a small amount of labeled data and a large amount of unlabeled data.\n",
    "- **What It Is**: This approach is useful when acquiring a large set of labeled data is difficult or expensive, but unlabeled data is abundant.\n",
    "- **Examples and Use Cases**: Semi-supervised learning is often used in situations where labeling data is costly or requires expert knowledge, such as in image and speech recognition tasks.\n",
    "- **Training Process**: The model makes predictions on the unlabeled data, then incorporates these predictions into its learning process, along with the labeled data.\n",
    "- **Combining Approaches**: It combines the strengths of both supervised and unsupervised learning, using the former to guide the learning process in the latter.\n",
    "\n",
    "\n",
    "#### What is Reinforcement Learning?\n",
    "Reinforcement Learning (RL) is a distinct type of Machine Learning where an agent learns to make decisions by performing actions and receiving feedback from these actions. It's heavily inspired by behavioral psychology.\n",
    "- **Basic Concept**: In Reinforcement Learning, an agent learns to make decisions by taking actions in an environment and receiving feedback in the form of rewards or penalties.\n",
    "- **What It Is**: Reinforcement learning is like learning by trial and error. The machine learns to make decisions by performing actions and receiving feedback.\n",
    "- **Examples and Use Cases**: Training a robot to navigate a maze, or developing an AI to play and win video games.\n",
    "- **Goal**: The primary goal in RL is to find a strategy, known as a policy, that decides what action the agent should take in a given state.\n",
    "- **Key Components**: RL involves agents, environments, states, actions, and rewards. The agent interacts with the environment, receives feedback, and learns to maximize cumulative reward over time.\n",
    "  - *Agent*: The learner or decision maker.\n",
    "  - *Environment*: Everything the agent interacts with.\n",
    "  - *State*: A representation of the current situation returned by the environment.\n",
    "  - *Action*: What the agent can do. The set of all possible actions is called the action space.\n",
    "  - *Reward*: A feedback from the environment to assess the action's effectiveness.\n",
    "- **Process**:\n",
    "  - *Interaction*: The agent interacts with the environment in discrete time steps.\n",
    "  - *Observation*: At each time step, the agent receives an observation (current state of the environment).\n",
    "  - *Action*: Based on this observation, the agent takes an action.\n",
    "  - *Reward and Next State*: As a consequence of the action, the agent receives a reward and finds itself in the next state.\n",
    "  - *Goal*: Through this process, the agent learns to make decisions that would maximize its rewards over time.\n",
    "- **Challenges in Reinforcement Learning**:\n",
    "  - *Trial and Error*: The agent must discover the best actions through trial and error, which can be a complex and time-consuming process.\n",
    "  - *Delayed Rewards*: Rewards may be delayed, making it challenging to understand which actions are truly beneficial in the long term.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2481493f-90a2-4537-8438-058a43f13477",
   "metadata": {},
   "source": [
    "isintelligence and real-world scenarios.\r\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3d78daab-bef7-4226-a292-99d0d0252f04",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}