I have this idea in a very vague form which I have not seen implemented yet but maybe someone can tell me if it is out there already or not. It might not be worth a patent as its just an addition but if something similar is patent I need to know who has it.
Its Basically a procedure to suggest related applications on App stores to users.
- For New users: We suggest nothing unless they have started using app store for 'n' days.
- 'n' is always greater than 2 to be used with any extrapolation algorithm.
- Device records each run of the Application installed through App stores with timestamp.
- An indexing service would run at certain time of day to update app index database by time units 't'. This can be customized per user or per device.
- Indexing service will place related tags for apps running at time 't' and using those tags would use same or different algorithm to find other applications.
- This can be used to just suggest user what apps he should be looking for at certain time of the day or extrapolate his next app installation from app stores.
Now in terms of why I came up with this idea is this:
I use my phone during the day to send and receive emails. I have to have an email widget on screen for this and it helps me. Between 9-5 this is all good, except at lunch time I want to use facebook and twitter. May be some devices are storing such patterns on phone or tablets. As per websites (I know google analytics does that and record user behaviour on web pages and can relate what time a visit was done however while its doing that, it does not highlight to user for example that 90% people who came to my question are likely to go to this link. This may become annoying ) but there is no such option to make this available or hide this feature if not required. Its about using this collected data for suggesting time based results.
I would like to use it for both pattern analysis of existing users and prediction for future for an individual or prediction of future application uses by other people.
Is this kind of thing too general to patent or some has already got this implemented on some devices?