Availability of compliance documentation is dependent on your configuration. Contact your DataRobot representative for more information. Please provide details you need in compliance document for unsupervised use case as this type of compliance is being under development and will be beneficial to provide it sooner.
As a user of DataRobot, I think that DataRobot does not provide a specific option to generate a compliance document for K-means clustering. However, you can create your own compliance document by following these general steps:
Understand the K-means clustering algorithm: Familiarize yourself with how the K-means clustering algorithm works, its assumptions, and its limitations. This is important to ensure that the document accurately explains the algorithm to non-technical stakeholders.
Document the data preparation process: Describe the steps you have taken to clean, preprocess, and transform your data before using it in the K-means clustering model. This includes handling missing values, scaling or normalizing data, and feature engineering.
Explain the choice of parameters: K-means clustering typically requires you to set the number of clusters (K) and the initialization method. Document the reasoning behind your choices, and mention any other relevant parameters used in your specific implementation.
Detail the model evaluation process: Explain how you assessed the quality of the clustering results, including any metrics used, such as silhouette scores, inertia, or other cluster evaluation techniques. If you tested multiple parameter values, mention the comparisons and results.
Describe the interpretation of results: Explain how the resulting clusters were analyzed, including any relevant visualizations, such as scatter plots or heatmaps, and how these clusters can be used in a business context.
Address data privacy and security concerns: Explain any steps taken to protect the privacy and security of the data used in the clustering process. This might include anonymization, data access control, and encryption.
Outline potential biases and limitations: Discuss any potential biases in the data or the clustering process, as well as any limitations of the K-means clustering algorithm in your specific use case.
Once you have created a comprehensive compliance document, you can share it with relevant stakeholders to demonstrate your understanding of the model, its results, and the potential implications of its use. It's important to keep the document up to date if any changes are made to the model or data used in the future.
For the most up-to-date information, I recommend reaching out to DataRobot directly, visiting their website, or subscribing to their newsletters and announcements. They should be able to provide you with the latest information on product updates and compliance document releases.
DataRobot continuously works to improve its platform and add new features, including compliance documents for different types of learning algorithms. While I don't have a specific release date for compliance documents related to unsupervised learning, I suggest reaching out to DataRobot support or your account manager for the most up-to-date information. They should be able to provide you with the latest updates on this feature's availability.