The application builds a neural network in the Python language.
Datasets are input, where:
- azimut - type: number, description: azimuth of phone sensor
- pitch - type: number, description: pitch of phone sensor
- roll - type: number, description: roll of phone sensor
- missClicks - type: number, description: user missed the button
- textDistance - type: number, description: average Levenshtein distance between words
- textTime - type: number, description: how long user retype original text in ms
- textErased - type: number, description: average number of symbols user erased
- leftCount - type: number, description: average number of taps for left hand
- rightCount - type: number, description: average number of taps for right hand
- state - type: number, description: state of health in time
- pill - type: number, description: number of tablets that user took in time
- dyskinesia - type: number, description: number of dyskinesias per time
- speed - type: number, description: average user speed per time
- volume - type: number, description: average user voice volume
- pauseCount - type: number, description: avarage number of pauses between words
- pauseTime - type: number, description: average pause time
Details: State has 3 meanings: good(1), bad(0) and neutral(0.5). The output should determine the patient's condition has improved, worsened or remained unchanged.