A company needs to classify human genes into 20 categories and requires an algorithm where the internal decision process can be documented and inspected. Which machine learning algorithm meets this requirement?

Decision trees are a suitable choice because they are inherently interpretable. Their tree-like structure explicitly shows the rules and features used at each decision point to classify data, making the internal decision process transparent and easy to document and inspect. This satisfies the requirement for an auditable algorithm. Linear regression is used for predicting continuous values, not for classification into categories. Logistic regression is a classification algorithm, but its internal workings, while simpler than neural networks, are less directly interpretable than the explicit rules of a decision tree. Neural networks are powerful for classification but are often considered "black boxes" due to their complex, non-linear internal structures, making their decision processes difficult to document and inspect.