Chapter 3 of 4 · Databricks Machine Learning Associate
Model Development
Algorithm choice, imbalance, Spark ML pipelines, Hyperopt tuning, cross validation, metrics.
- 5topics
- 25 minfast track
- 2diagrams and exhibits
- 31%of the exam
Topics in this chapter
Take them in order, or jump to the one you need.
- 1Algorithms, Imbalanced Data and Bias VarianceDatabricks ML Associate: pick an algorithm for the scenario, fix imbalanced training data with sampling or weights, read bias and…5 min
- 2Estimators, Transformers and Training PipelinesDatabricks ML Associate: tell Spark ML estimators from transformers, know what fit and transform return, and chain the stages in…Diagram5 min
- 3Hyperparameter Tuning with HyperoptDatabricks ML Associate: tune with Hyperopt fmin, compare random, grid and Bayesian search, and parallelize single node models…5 min
- 4Cross Validation and Grid SearchDatabricks ML Associate: cross validation vs a train validation split, CrossValidator in Spark ML, and counting the models a grid…Diagram5 min
- 5Classification and Regression MetricsDatabricks ML Associate: use F1, log loss, ROC AUC, RMSE, MAE and R squared, pick the metric for the goal, and exponentiate log…5 min
About this chapter
What this chapter covers
Model Development is 31% of the Databricks Machine Learning Associate exam, the second largest chapter. It covers how you turn prepared data into a trained, tuned and fairly evaluated model.
The chapter has five parts: choosing an algorithm and handling imbalanced data and model complexity, building Spark ML pipelines from estimators and transformers, tuning hyperparameters with Hyperopt, validating with cross-validation and counting the models it trains, and picking the right classification or regression metric.
Questions are single-answer and multiple-selection ("Which TWO ...?") questions, mostly short scenarios. Some show a short Python snippet (MLflow, Feature Engineering client, Spark ML, Model Serving) and ask what it does or which call meets a requirement. There are no exhibits.
How the topics fit together
You first choose an algorithm that suits the label and the data, then wrap the preprocessing and the model in one pipeline. Tuning searches the hyperparameters, cross-validation scores each setting fairly, and the metric you choose decides which model wins.
- Algorithm questions ask which model type fits a scenario, how to handle a rare class and whether a model is overfitting.
- Pipeline questions ask whether a stage is an estimator or a transformer and what fit() returns.
- Tuning questions ask what fmin needs, which search algorithm to use and when SparkTrials helps.
- Validation questions ask for the number of models trained and the trade-off of cross-validation.
- Metric questions ask which metric fits the business goal and how to read it on a log scale.
Count before you answer. Many questions in this chapter hide a small calculation: models trained by a grid with k folds, F1 from precision and recall, or an error that only makes sense after exponentiating a log prediction.
How to use these pages
Each topic page has the same shape, so you can skim for what you need.
- What it is: the idea in plain words.
- How it works: the mechanism, with tables and examples.
- What to remember for the exam: the facts to memorize and how questions are phrased.
- Common mistakes: the classic traps.
- Check yourself: three questions to test your understanding.
Ready for more practice?
The full Databricks Machine Learning Associate question bank has 600 questions, with timed practice exams and an explanation for every answer.
This study guide is independent and is not affiliated with, endorsed by, or sponsored by Databricks. Vendor names, certification names and exam codes are trademarks of their respective owners and are used for identification only.