Machine Learning Terms

Black box

A model or system whose internal workings are not easily interpretable or understood by humans, even though it may produce accurate or useful predictions.

---

Features and Labels

---

Features

• Features are input data (X-Features, Input Variable, Colums, etc...)

• Features are represented as vector.

$x=(x1, x2, x3,... xn)$

x: Feature & x1,x2,x3,...xn are Features of data (input)

---

Label

• Label is our answer or output (Y-Feature, Target Variable, etc...).

• Label is represented as

$y=f(x)$

y: Label and x: Feature

---

Model

In machine learning, a model is a system or an algorithm that learns patterns from data and uses those patterns to make predictions or decisions on new, unseen data.

---

Training a model

---

Training Face

Training a model is the process of teaching a machine learning model to learn patterns from a large dataset

---

Interfence Face

After training, the model is called a trained model. The process of using this trained model to make predictions on unseen data is called the inference phase.

---

Underfitting

Underfitting is a state of a machine learning model during training where the model cannot find patterns in the training dataset. Therefore, it cannot make accurate predictions and its performance is low.

Reason

• Model is too short.

• Inefficient training data.

---

Overfitting

Overfitting is a state of machine learning when training a complex model which faces a challenge while finding a pattern within the dataset along with noise. It will work perfectly while training data but poorly with unseen data.

---

Training Datasets

A training dataset consists of input features and their corresponding labels, which are used to teach a machine learning model during training.