NumPy with machine learning libraries - Step-by-Step Execution

Concept Flow - NumPy with machine learning libraries

Create NumPy array

↓

Pass array to ML library

↓

ML library processes data

↓

Get model output or result

↓

Use output for prediction or analysis

Start with a NumPy array, pass it to a machine learning library, process it, and get results for predictions or analysis.

Execution Sample

NumPy

import numpy as np
from sklearn.linear_model import LinearRegression

X = np.array([[1], [2], [3], [4]])
y = np.array([2, 4, 6, 8])
model = LinearRegression().fit(X, y)
pred = model.predict(np.array([[5]]))

Create NumPy arrays for features and target, train a linear regression model, and predict a new value.

Execution Table

Step	Action	Variable/Function	Value/Result
1	Create feature array X	X	[[1], [2], [3], [4]]
2	Create target array y	y	[2, 4, 6, 8]
3	Initialize LinearRegression model	model	LinearRegression object
4	Fit model with X and y	model.fit(X, y)	Model trained with coefficients
5	Predict with new input [[5]]	model.predict([[5]])	[10]

💡 Prediction step completed, output is the predicted value for input 5.

Variable Tracker

Variable	Start	After Step 1	After Step 2	After Step 4	After Step 5
X	None	[[1], [2], [3], [4]]	[[1], [2], [3], [4]]	[[1], [2], [3], [4]]	[[1], [2], [3], [4]]
y	None	None	[2, 4, 6, 8]	[2, 4, 6, 8]	[2, 4, 6, 8]
model	None	LinearRegression object	LinearRegression object	Trained model with coef_=2	Trained model with coef_=2
pred	None	None	None	None	[10]

Key Moments - 3 Insights

Why do we use NumPy arrays instead of lists with machine learning libraries?

Machine learning libraries like scikit-learn expect NumPy arrays because they are efficient and support mathematical operations needed for training models, as shown in steps 1 and 2 where arrays X and y are created.

What happens when we call model.fit(X, y)?

The model learns the relationship between X and y by calculating coefficients, as shown in step 4 where the model becomes trained with coefficients.

Why do we pass a 2D array to model.predict even for a single input?

The model expects input features in 2D shape (samples, features), so even a single input like [[5]] must be 2D, matching the shape of X used in training (step 5).

Visual Quiz - 3 Questions

Look at the execution table, what is the value of 'pred' after step 5?

A [8]

B [10]

C [5]

D [2]

💡 Hint

Check the 'Value/Result' column for step 5 in the execution table.

At which step does the model learn the coefficients?

A Step 3

B Step 2

C Step 4

D Step 5

💡 Hint

Look for the step where 'model.fit(X, y)' is called in the execution table.

If we pass a list instead of a NumPy array to model.fit, what would likely happen?

A It works the same without error

B Model raises an error or warning

C Model trains but with wrong results

D Model ignores the input

💡 Hint

Machine learning libraries expect NumPy arrays as shown in variable_tracker for X and y.

Concept Snapshot

NumPy arrays are the standard input for machine learning libraries.
Create arrays for features and targets.
Fit models using .fit(X, y).
Predict new data with .predict(new_X).
Always keep input shapes consistent (2D for features).

Full Transcript

This example shows how to use NumPy arrays with a machine learning library, scikit-learn. First, we create feature and target arrays using NumPy. Then, we initialize a LinearRegression model and train it with the arrays. After training, we predict a new value by passing a 2D array. The execution table traces each step, showing variable values and actions. Key moments clarify why arrays are needed, what fitting does, and why input shape matters. The visual quiz tests understanding of prediction output, training step, and input types. The snapshot summarizes the process for quick reference.