Skip to content
Neural Soft
Contact us Español (View this page in Spanish)

Data, machine learning and deep learning

Machine learning with Python course

Real predictive models, from the data to evaluation and interpretation.

Request information

Machine learning is the foundation of most non-generative AI applications: forecasting demand, detecting fraud, anticipating customer churn or segmenting the market. And in many cases it is still the most effective and cheapest tool for problems with structured data.

This course covers the full cycle with Python and scikit-learn: framing the problem, exploring and preparing the data, training and comparing models, validating without fooling yourself and explaining predictions. With a project on a business case from start to finish.

It is the foundation for the deep learning with PyTorch course.

What you will learn

  • Frame a business problem as a machine learning problem
  • Prepare data and create useful features
  • Train and compare regression and classification models
  • Validate properly and choose the right metric
  • Interpret models and explain their predictions

Syllabus

9 modules, from the fundamentals to the final project.

  1. What machine learning is

    • Supervised, unsupervised and reinforcement learning
    • From business problem to ML problem
    • When machine learning is not needed

  2. Exploratory analysis

    • Distributions, correlations and outliers
    • Data quality

  3. Data preparation

    • Missing values and categorical variables
    • Scaling and feature engineering
    • scikit-learn pipelines

  4. Regression

    • Linear regression and regularisation
    • Error metrics

  5. Classification

    • Logistic regression, trees and random forests
    • Gradient boosting (XGBoost, LightGBM)
    • Imbalanced classes

  6. Validation and metrics

    • Cross-validation and data leakage
    • Precision, recall, F1, AUC and confusion matrix
    • Hyperparameter tuning

  7. Unsupervised learning

    • Clustering with k-means and DBSCAN
    • Dimensionality reduction
    • Customer segmentation

  8. Interpretation and fairness

    • Feature importance and SHAP values
    • Bias in data and models

  9. From notebook to real use

    • Saving and serving a model
    • Monitoring performance

Tools you will practise with

  • Python
  • scikit-learn
  • pandas
  • XGBoost or LightGBM
  • SHAP
  • Jupyter

Hands-on project

You develop a complete predictive model for a business case (for example, predicting customer churn or late payments), from exploratory analysis to validation, interpretation and a presentation of results for the business.

Requirements

  • Basic Python (equivalent to the Python for AI course).
  • Some statistics recommended.

Other courses

Where it is applied

  • Custom AI development

    We design and build artificial intelligence solutions tailored to your processes, data and tools when an off-the-shelf solution does not fit or falls short.

  • AI customer churn prediction

    A machine learning model spots which customers are most likely to leave and why, so the sales team can act before losing them.

  • AI predictive maintenance

    Sensor data from your machines warns of wear before a breakdown, so repairs happen during a planned stop rather than in the middle of production.

Want this course for your team?

We adapt the content, length and format to your company.

Request information