Assessing diseases by analyzing gait measurements

A system comprising:

• non-transitory computer readable storage media that stores gait data comprising:
• training data indicative of abnormal gaits and normal gaits;
• and
• sensor data indicative of a user gait;
• a hardware computer processor that:
• stores a neural network that includes a bidirectional recurrent neural network layer, an attention layer, and a fully connected layer:
• parses the training data into a series of training data windows;
• trains the recurrent neural network layer to bidirectionally analyze each training data window in series and generate two context vectors characterizing each training data window and the previously analyzed training data windows;
• trains the attention layer to determine a relative importance of one or more of the context vectors for distinguishing between training data indicative of normal user gaits and training data indicative of abnormal user gaits;
• weights one or more of the context vectors based on the relative importance of each context vector;
• trains the fully connected layer to distinguish between normal gaits and abnormal gaits based on one or more of the weighted context vectors characterizing the training data;
• parses the sensor data into a series of sensor data windows;
• bidirectionally analyzes each sensor data window in series, by the recurrent neural network layer, to generate two context vectors characterizing each sensor data window and the previously analyzed sensor data windows, each context vector representing an interim assessment of the normality of the user gait;
• weighs one or more of the context vectors characterizing the sensor data based on a relative importance of each context vector;
• and
• generates a final assessment characterizing the user gait as normal or abnormal, by the fully connected layer that has been trained using the training data, using one or more weighted context vectors characterizing the sensor data.