7.1. What is Modelling?

Modelling refers to the process of: • Collecting data • Analyzing patterns • Building a logic or mathematical structure • Training the machine to recognize or predict outcomes

Modelling in AI is a systematic approach comprising several steps that work together to teach machines. First, data is collected, which serves as the foundation for everything else. Next, patterns within the data are analyzed to understand how different factors influence each other. Following that, a structure—often mathematical or logical—is crafted based on these patterns. Finally, the machine is trained using this structure so that it can recognize similar patterns in new data or make accurate predictions.

Think of modelling like a coach training a football team. The coach gathers various data about the players’ performances (collecting data), analyzes their strengths and weaknesses (analyzing patterns), and devises a strategy to improve their skills (building a logic or mathematical structure). The training sessions that follow help the players learn this strategy so they can perform better in matches (training the machine).

In simple terms, it is like training a student using various examples (data), so they can solve similar problems on their own later (prediction or classification).

When we model in AI, the approach can be likened to teaching a student. Educators provide students with numerous examples and exercises, which help them grasp concepts or solve problems. Similarly, a machine learns from provided data, which serves as examples of the relationships and outcomes it needs to understand. Once trained, just as a student can apply what they've learned to new questions, the machine can apply its training to new data.

Imagine a student learning mathematics. The teacher first provides several examples of addition problems and explains the methods to solve them. After enough practice, the student can solve new addition problems independently. Similarly, a machine learns to identify whether a fruit is an apple or an orange by being shown many examples of both, allowing it to independently classify new fruits later.

Suppose you want to build a model to identify whether a fruit is an apple or an orange. You provide the machine with features (data) like: • Color • Weight • Shape Using this data, the machine learns how apples differ from oranges and can later predict new fruits correctly.

In this example, we want to train a model to differentiate between apples and oranges. We collect specific features about these fruits, such as their color, weight, and shape. The model analyzes this data to understand the common characteristics of each fruit. After training on this data, the machine learns the distinguishing features, enabling it to accurately classify new fruits it encounters based on the knowledge it gained.

Consider a fruit vendor who sorts apples and oranges. The vendor notices that apples are generally red or green, heavier, and round, while oranges are typically orange, lighter, and spherical. Over time, the vendor becomes adept at identifying fruits just by looking at them. Similarly, a model learns these characteristics to classify unknown fruits reliably.

--