Notification Diary uses machine learning to automatically handle incoming notifications - hide unwanted and unimportant ones, and show notifications with valuable information. Notification Diary learns from your interactions with the notifications, and your provided ratings of the importance of the notification contents and the situation where the notifications were presented to you.
The application requires a short setup period after installation, and short periods of user-given feedback and training during use.
Based on your provided information as training data, Notification Diary uses machine learning to predict whether a notification should be shown or hidden from you.
This will reduce the amount of burden of constant disruptions from your smartphone.
You can verify the accuracy of each prediction and continuously retrain the application to understand your preferences in detail.
Notification Diary was created for research purposes and part of data collected will be anonymously stored for research. Notification Diary collects the following anonymised information from your device:
Screen state (on/off), Network availability, WiFI availability, anonymised location tags ("location a", "location b"), foreground application package names, battery level, ringer mode, and headphone connectivity. It also monitors for incoming calls but does not store calls or any information during a call.
This information is periodically uploaded to our secure server. We do not share any of the data collected with third parties and the collected data is only used by us for our research purposes. You can find examples of our research at http://ubicomp.oulu.fi
Sensitive information, e.g. notification contents (text) is only stored locally on your device and never uploaded to our server.
For any questions regarding the application, please contact the author.