Skip to main content

Tutorials

Core​

  • Tutorial 1A: Introduction to H2O Driverless AI

    This tutorial is a quick start of H2O Driverless AI (DAI) and gives you a quick overview of the high-level capabilities and use cases you can achieve using H2O Driverless AI.

    We will explore the Titanic dataset from the perspective of a passenger life insurance company and analyze the possible risk factors derived from this dataset that could have been considered when selling passenger insurances. More specifically, we will create a predictive model to determine what factors contributed to a passenger surviving.

  • Tutorial 1B: Machine learning experiment scoring and analysis - financial focus

    Tools like ROC and Precision-Recall curves are used to assess how well classification models predict outcomes. In this tutorial, you will use a subset of the Freddie Mac Single-Family Loan-Level dataset to build a classification model and use it to predict if a loan will become delinquent. Through H2O Driverless AI's Diagnostic tool, you will examine the financial impacts the false positive and false negative predictions have while exploring tools like ROC Curve, Prec-Recall, Gain Charts, Lift Charts, and K-S Chart. Finally, we'll explore several metrics like AUC, F-Scores, GINI, MCC, and Log Loss to evaluate the generated model's performance.

  • Tutorial 1C: Building and interpreting a machine learning model with Python for credit default prediction

    In this tutorial, we will explore how to automate the process of building, analyzing, and interpreting machine learning models using H2O Driverless AI's Python API. This hands-on guide focuses on predicting credit default risk using the UCI Credit Card dataset, a popular dataset in financial modeling.

Time series & NLP​

  • Tutorial 2A: Building, interpreting, and scoring time series models with H2O Driverless AI

    In this tutorial, we will learn how to build, interpret, and score a time series model in H2O Driverless AI.

    A time series model is a predictive model that leverages machine learning algorithms to analyze and forecast data points in a temporal sequence. Unlike traditional statistical models, time series models can capture complex patterns and relationships within the data by learning from large amounts of historical data. These models can handle non-linear patterns and interactions between variables and even incorporate external data sources.

    Time series models are utilized in various domains, such as finance (stock price prediction), healthcare (patient monitoring), retail (sales forecasting), and more. They offer the advantage of handling complex and high-dimensional data for more accurate and robust predictions.

    Throughout the tutorial, we will harness a synthesized version of the renowned dataset, Walmart Recruiting—Store Sales Forecasting, to construct a time series regression model that predicts the sales for several departments in 45 Walmart stores in different regions. We will also learn how to generate a machine learning interpretability (MLI) report for a time series model in H2O Driverless AI.

    The MLI report is particularly valuable for enhancing machine learning interpretability, especially in regulated industries where transparency and explanation are crucial. With a model's MLI report, you can uncover valuable insights into its inner workings, such as feature importance and other performance metrics. This report can help you better understand and trust the predictions made by the model, offering deeper insights into your built model and enabling you to make informed decisions in machine learning applications with the model.

  • Tutorial 2B: Natural Language Processing (NLP) - Sentiment analysis

    This tutorial describes how to launch a sentiment analysis experiment, including the experiment settings and custom recipe, and concludes with a challenge to test your skills.

Model deployment​

  • Tutorial 3A: Build, deploy, score, and monitor an H2O Driverless AI model with H2O Driverless AI and H2O MLOps

    In this tutorial, you'll explore how to build, deploy, score, and monitor an H2O Driverless AI model with H2O Driverless AI and H2O MLOps.

    Throughout the tutorial, we'll harness the renowned Default of Credit Card Clients Dataset to construct a classification model predicting clients' probability of defaulting on their next credit card payment. Unlike previous tutorials, right after building an H2O Driverless AI model, our primary focus will be understanding how to deploy, score, and monitor an H2O Driverless AI model with H2O MLOps.

    Understanding how H2O Driverless AI and H2O MLOps can streamline the process of building to monitoring a model can empower you to use an end-to-end solution that automates and optimizes the entire machine learning workflow—from model development to deployment and ongoing monitoring—ensuring that models are not only built efficiently but also maintained effectively in production.

Use cases in banking sector​

  • Tutorial 4A: Building a regression model for loan prediction with H2O Driverless AI

    In this tutorial, you'll explore how to build a regression model for loan prediction with H2O Driverless AI.

  • Tutorial 4B: Building a classification model for credit card clients information with H2O Driverless AI

    In this tutorial, you'll explore machine learning interpretability (MLI), an essential component as industries from healthcare to banking increasingly adopt machine learning models for various applications. With predictions influencing crucial decisions such as healthcare costs and loan approvals, the interpretability of these models becomes paramount to ensure regulatory compliance, prevent discrimination, and guard against adversarial attacks.

    Throughout the tutorial, we'll harness the renowned Default of Credit Card Clients Dataset to construct a classification model predicting clients' probability of defaulting on their next credit card payment. Unlike previous tutorials, our focus will be on understanding and interpreting a machine learning model rather than the experimentation process itself. We'll explore metrics and visualizations generated by a model's MLI (machine learning interpretability) report.

    This report is particularly valuable for enhancing machine learning interpretability, especially in regulated industries where transparency and explanation are crucial. With a model's MLI report, you can uncover valuable insights into its inner workings, such as feature importance, partial dependence plots, and other performance metrics. This process can empower you to understand better and trust the predictions made by the model while offering you deeper insights into your built model and empowering you to make informed decisions in machine learning applications.

  • Tutorial 4C: Building an NLP model for customer complaints with H2O Driverless AI

    In this tutorial, we will explore how to use H2O Driverless AI to build a Natural Language Processing (NLP) model for analyzing customer complaints in the banking sector. We will use the dataset from Kaggle's Consumer Complaint Resolution Dataset, with the target column Customer disputed?. This target indicates whether a customer disputed the resolution of their complaint.

  • Tutorial 4D: Building a time series model for NPS with H2O Driverless AI

    Understanding and predicting NPS over time allows banks to take proactive steps to address customer concerns, improve services, and boost overall satisfaction. This tutorial demonstrates how to build a time series model using H2O Driverless AI to forecast NPS, enabling data-driven decision-making in the banking sector.

Feedback