caregiver_recomm

Usage

from caregiver_recomm import Recommender

recomm = Recommender()
recomm.dispatch(speaker_id, evt)

Recommender(evt_dim=4)

• evt_dim: Event dimensions, default 4
• mock: Simulate the user interaction, default False
• server_config: Remote server configuration, default None
  • client_id: client id
  • url: server url

dispatch(speaker_id, evt)

• speacker_id:
• evt: numpy array or python list

EMA Tables Used

• reward_data: Information of all prompts sent to the phone are stored in this table This table holds the following: speakerID, empathid,TimeSent,suid,TimeReceived,Response,Question,QuestionType,QuestionName,Reactive, SentTimes,ConnectionError,Uploaded
• ema_storing_data: Information pertaining only to recommendation messages sent are stored in this table This table holds the following: time,event_vct,stats_vct,action,reward,action_vct,message_name,uploaded
• recomm_saved_memory: One row containing the unique deployment home id, baseline period time left in seconds, the most recent time the baseline period left was updated, morning start time, evening end time, and max messages. This table holds the following: deploymentID, baselineTimeLeft, lastUpdated, morningStartTime, eveningEndTime, maxMessages All are default to 'NA' except for maxMessages which defaults to 4
• ema_settings: Used to dynamically change prompts and answer choices
• ema_data: Used to retrieve prompt answer