Determining optimal machine learning models

One or more non-transitory computer-readable media storing instructions that, when executed by a computing platform comprising at least one processor, and memory, cause the computing platform to:

• identify one or more initial features indicative of login activity to an enterprise resource;
• generate, based on the one or more initial features, an initial test dataset;
• generate, based on the initial test dataset, a test model;
• deploy the test model in a production environment;
• generate a feature engineering table, wherein a row of the table represents a feature of the one or more initial features, and a column represents a candidate training model, and an entry in a cell corresponding to the row and the column is indicative of a significance of the feature to the candidate training model;
• identify, based on the deploying and the feature engineering table, one or more production feature vectors;
• generate, based on the one or more production feature vectors, a training dataset;
• perform, for the training dataset and based on one or more performance criteria, a data quality check;
• predict, by applying Bayesian optimization to the training dataset, an optimal machine learning model;
• and
• apply the optimal machine learning model to detect unauthorized attempts to access the enterprise resource.

A method, comprising:

• at a computing platform comprising at least one processor, and memory:
• collecting, via a network, data indicative of login activity to an enterprise resource;
• identifying one or more initial features indicative of the login activity;
• generating, based on the one or more initial features, an initial test dataset;
• generating, based on the initial test dataset, a test model;
• deploying the test model in a production environment;
• generating a feature engineering table, wherein a row of the table represents a feature of the one or more initial features, and a column represents a candidate training model, and an entry in a cell corresponding to the row and the column is indicative of a significance of the feature to the candidate training model;
• identifying, based on the deploying and the feature engineering table, one or more production feature vectors;
• generating, based on the one or more production feature vectors, a training dataset;
• performing, for the training dataset, a data quality check;
• and
• determining, by applying Bayesian optimization to the training dataset, an optimal machine learning model.