# Is there any K-means analog for regression, not culstering?

Good day Is there any regression Machine learning model, based on k-means? Example (Zillow's Zestimate): I have a lot of houses(with features like sq.feet and location) and prices. I want to estimate price for a new house by averaging k neighbors.

You could use $k$-nearest neighbors regression. It's not based on $k$-means, but it is based on estimating the price of a new house by averaging $k$ neighbors (in particular, the $k$ most "similar" houses in the training set, for some measure of similarity that you define). $k$-nearest neighbors is supported in many machine learning toolkits.