Pluralsightbeginneren

Building Features from Numeric Data

Master data preprocessing with scikit-learn! Learn to scale, normalize, and transform data for optimal machine learning model performance. Build powerful features.

4.8(24)
2h 25m
Certificate
Janani Ravi
Janani RaviInstructor

Course Content

Building Features from Numeric Data is a beginner-level Pluralsight course taught by Janani Ravi. It runs 2h 25m, is taught in English, and includes a certificate of completion. Learners rate it 4.8 out of 5 from 24 reviews. It is a paid course, first published April 2019.

Enhance Machine Learning Models with Effective Data Preprocessing

The quality of numeric data preprocessing is crucial for successful machine learning. Optimized data pre-processing significantly accelerates model training and validation, saving you valuable time and resources while dramatically improving prediction accuracy. This course, Building Features from Numeric Data, equips you with the skills to design and implement robust, mathematically sound data pre-processing pipelines.

Key Topics Covered:

  • Normalization, Standardization, and Scaling: Understand the fundamental importance of these techniques and learn how to effectively adjust the central tendency and dispersion of data features.
  • Outlier and Anomaly Detection: Learn practical methods for identifying and handling outliers and potentially anomalous data points within your datasets.
  • Scaling and Normalization Techniques: Master essential scaling and normalization methods, including:
    • L1-norm, L2-norm, and Max Norm Normalization: Transform feature vectors to achieve uniform magnitude – widely used in applications like calculating cosine similarity for document vectors and image preprocessing for convolutional neural networks (CNNs).
  • Data Transformation: Explore advanced data transformations, including:
    • Quantization: Learn how to reduce the number of distinct values in your data.
    • Custom Transformers: Develop bespoke transformation methods tailored to specific use cases.
  • Log and Power Transformations: Discover how to apply log and power transformations to your data.
  • Non-Normal Data Handling: Compare and contrast the effectiveness of:
    • Yeo-Johnson Transform
    • Box-Cox Transform
    • Quantile Transformation
      in transforming data with non-normal distributions (e.g., chi-squared, lognormal) into a bell curve shape – a common requirement for optimal model performance.

By the end of this course, you will possess the essential data preprocessing and transformation skills needed to maximize the performance of your machine learning models. Learn how to prepare your numeric data for optimal results and unlock the full potential of your ML projects.

$ whoami — the devs behind tutorialsearch.io

Want AI in your product? Hire the team behind TutorialSearch.io

LLM features, RAG, agents and automation wired into a real product — built by the devs who shipped this site.

5.0
2h 42m
intermediate
pluralsight
5.0(28)
1h 31m
all
skillshare
5.0
1h 28m
beginner
pluralsight