Understand machine learning

i am facing difficulty in understanding the concepts of machine learning.it would be very heplful if someone explains it.

1 Like

Hi @Shivansh_Sharma , Thanks for using Discourse. You could see the video on YouTube Video by Fireship that explains it in simple words to alleviate your difficulty.
Or you could see this video by Simplilearn. (Machine Learning | What Is Machine Learning? | Introduction To Machine Learning | 2021 | Simplilearn - YouTube)

If you prefer reading, here is a brief introduction to Machine Learning that might help you →

Machine learning is a subfield of artificial intelligence (AI) that focuses on developing algorithms and models that enable computers to learn and make predictions or decisions without being explicitly programmed for each specific task.

Terminology →

  1. Data: Data is at the heart of machine learning. Based on the information in the data, machine learning algorithms identify patterns and create predictions. Data can take many forms, including text, graphics, and numerical numbers.

  2. Training: During the training phase, a labeled dataset is delivered to a machine learning model. This dataset is made up of input data (features) and desired outputs (labels or targets). The labeled data is used by the model to find patterns or correlations between the input features and the desired outputs.

  3. Features and Labels: Features are the observable features or attributes of the data that are used to produce predictions. Labels are the descriptive terms for the data. Labels are the desired outcomes or values that the model tries to predict or classify based on the input information. They are also known as targets or outputs.

  4. Machine learning model: A machine learning model is a mathematical representation or algorithm that has been trained on labeled data. It extracts patterns, connections, or statistical qualities from data and applies them to generate predictions or choices based on previously unseen data.

  5. Prediction: Once trained, the model can be used to generate predictions or judgments based on fresh, previously unknown data. The model uses the fresh data’s input attributes and learned patterns to make predictions or classifications.

  6. Evaluation: The machine learning model’s performance is evaluated by comparing its predictions or judgments to a distinct set of data known as the test set. The model’s performance is measured using metrics such as accuracy, precision, recall, and mean squared error.

  7. Iteration and Improvement: Machine learning is a process that is iterative. If the model’s performance is not adequate, changes can be done to increase its accuracy or efficacy, such as selecting other characteristics, improving the model architecture, or modifying its parameters.

Types of Machine Learning using Analogies →

  1. Supervised Learning: Assume you have a cake-baking recipe with a list of ingredients and step-by-step directions. The recipe serves as the labeled data in this example. You carefully follow the directions, and the result is a lovely cake. Similarly to supervised learning, the recipe gave clear direction throughout the process. The labeled data, often known as the recipe, assists the model (you) in understanding the relationships between the components (features) and the desired outcome (the cake).

  2. Unsupervised Learning: Assume you are given a variety of ingredients but are not given a recipe or any instructions. Your objective is to look for patterns or groups of related elements. You begin studying the ingredients based on their qualities, such as colour, texture, or taste, without any prior knowledge. Through this investigation, you discover that some elements, such as fruits, belong to one group while others, such as dairy products, belong to another. Similar to unsupervised learning, you learned the basic structure or patterns of the ingredients without any explicit direction. This sort of learning identifies hidden patterns or structures in data.

  3. Reinforcement Learning: Assume you want to make a one-of-a-kind cake recipe but are unsure about the proper measurements and ingredients. To begin, try varying amounts of flour, sugar, eggs, and other ingredients. You taste the cake and receive comments after each attempt. If the cake tastes nice, it validates the decisions you made in that attempt. If it doesn’t taste right, you change the ingredients and try again. You progressively learn which combinations of components result in a wonderful cake through this trial-and-error approach. Iterative learning is related to reinforcement learning in that you receive feedback (positive or negative) based on your actions. The goal is to maximize the reward (a delicious cake) by making the right choices (choosing the right amount of ingredients).

Hope this answers your query, and please feel free to ask further queries.

1 Like


1 Like