Hello. I have historical data of real estate. Includes attributes like price, revenue, maintenance, maintenance debt etc. I had two ideas for the data previously, to predict future real estate price and to use some clustering algorithm to put the properties into categories (good, average bad or something like that). Now I got this idea to generate clusters for any given point in time and use that history of clusters to predict migration of properties between clusters.
I am aware of inter cluster migration estimation but that seems to be a prediction of how points shift inside a cluster with the introduction of new data points rather than a timeline of the movement of points in the cluster.
Writing this I’m also thinking it might be possible to simply treat the history of clusters as a category to be predicted e.g. given the history of property X that is in category A the prediction is that in 6 months it’s heading for cluster B.
Does anyone have experience with similar work? Are there papers I am missing?
Thanks in advance.