Figure 1 | Scientific Reports

Figure 1

From: Predictive modeling of depression and anxiety using electronic health records and a novel machine learning approach with artificial intelligence

Figure 1

This is the pipeline used to train the machine learning models and generate predictions. The training set is sent through fivefold training for each model type to generate a prediction for each training sample. These predictions are then used to train a higher level model to predict a final outcome given the predictions from the fivefold training. Each of the 6 models from each fold then predicts on the held out test set and the average prediction for the probability of depression is stored. The higher level model then makes final predictions on the held out test set.

Back to article page