sentence_classification

Example 1 - Sentence Classification - ADE Detection

This example highlight the detections of adverse drug events (ADE). Each sentence is classified for the absence (not relevant) or the presence (relevant) ADE.

Model Set Up

Model Initializing

The sentence_classification\resources folder contains several files required for the model (not tracked by GitHub).

In the __init__ method of the model class, the model is loaded using tensorflow. Additionally, proprietary classes load the model configuration and the embeddings of the model.

Pre-processing of inputs

In the preprocess_model_input method, the input text is converted into a numerical representation using word embeddings.

Prediction of inputs

The predict method is a simple wrapper for the predict method of the tensorflow method of the model.

Post-processing of model outputs

The output of the model is a numpy array representing the label and the confidence numerically. Numpy is used to extract the label and the confidence, which are returned to the Annotator class.

Sentence classification annotations

Exemplary annotation for an input sentence:

annotation = {
                "begin": 0,
                'end': 25,
                "value": relevant,
                'label': 'ade-sentence-classification',
                'confidence': 0.68,
                'componentId': 'ade-sentence-classifier:0.1.0'
            }