proper sequoia
i want to make a platform like taskrabbit, how do i make a machine learning recommendation system based on nearest location and tasker preferred task category. in the map it will show all nearest and tasker preferred category
if it does not possible what other to be applied to this platform?
proper sequoia
i try to use knn but i think it doesnt good for this task, i dont know i need a view from others
proper sequoia
i know maybe there is a algorithm without ml to do that, so thats why i asking for other idea to be applied with ml
hollow phoenix
Do you have enough taskers for this to be a problem?
If you need to find the people closer to a location you can look into GIS, for example https://postgis.net/ or even https://www.gaia-gis.it/fossil/libspatialite/index
proper sequoia
Do you have enough taskers for this to be a problem?
The ratio between tasker and task will be 1:10 so it let the tasker decide which taks that they will pick
I mean is this can be solve as a machine learning task? Because i need to do a machine learning task for a capstone
hollow phoenix
Ah, this is a capstone project. I see.
this is the state-of-the-art in terms of multi-feature streams for recommendation: https://medium.com/nvidia-merlin/recommender-systems-not-just-recommender-models-485c161c755e
so you can incorporate location at the retrieval stage
proper sequoia
Filtering: what tasker like todo based on category
Scoring: based on nearest location?
I dont know about the ordering, but is this correct?