Scaling Features with StandardScaler

In today's lesson, you'll learn how to standardize financial data using the StandardScaler from the sklearn library. Scaling features ensure that all data contribute equally to machine learning models, improving their performance and robustness.

Lesson Goal: By the end of this lesson, you will be able to effectively scale financial features and understand the importance of this step in preparing data for machine learning.

Let's quickly recall how to load and preprocess the Tesla stock dataset:

import pandas as pd
import datasets

# Load the dataset
data = datasets.load_dataset('codesignal/tsla-historic-prices')
tesla_df = pd.DataFrame(data['train'])

# Feature Engineering: creating new features
tesla_df['High-Low'] = tesla_df['High'] - tesla_df['Low']
tesla_df['Price-Open'] = tesla_df['Close'] - tesla_df['Open']

We've successfully loaded the Tesla dataset and created new features: High-Low and Price-Open.

Feature scaling is crucial for machine learning for several reasons:

• Equal Contribution: Ensures all features contribute equally to the model.
• Improved Convergence: Helps in faster convergence during model training by making gradients less sensitive to feature magnitude.
• Prevent Dominance: Prevents features with larger scales from dominating those with smaller scales.

Feature scaling is particularly useful in scenarios like:

• Predicting House Prices: Square footage in thousands vs. the number of bedrooms in single digits.
• Stock Market Analysis: Stock price in hundreds vs. trading volume in millions.
• Health Data: Age in the 0-100 range vs. blood pressure in the hundreds.
• Retail Sales Prediction: Number of items sold vs. store rating in single digits.

These examples highlight the importance of scaling to ensure uniform treatment of features, thereby enhancing model performance.

Standardization involves transforming your data so that the mean of each feature is 0 and the standard deviation is 1. This process ensures all features on the same scale, improving the performance and robustness of machine learning models. The formula for standardization is:

$z = \frac{(x - \mu)}{\sigma}$

where:

• $z$ is the standardized value,
• $x$ is the original value,
• $\mu$ is the mean of the feature, calculated as the average of all values of that feature,
• $\sigma$ is the standard deviation of the feature, which measures the amount of variation or dispersion of the values.

By applying this formula, each feature will have a mean of 0 and a standard deviation of 1, enabling more stable and faster convergence during the training of machine learning models.