Hello, all participants!
According to the rules, the code sharing is forbidden, but we can discuss approaches, isn’t it? Especially since the deadline has passed.
Lets talk about your approachs in Snowcast Snowdown competition
1.What data you considered ? What datasets are helpful for your models?
2.What machine learning models you are used ?
3.Models stacking and blending
4.Training and validation
1.I used regions, stations types, DEM, GlobCover, soil map and month averaged MODIS. Also take into account the last 5 SNOTEL/CDEC measurements.
2.My solution is optimal interpolation based on neural network gaussian process.
In this approach the parameters of the gaussian process calculation using the neural network. The correlation function is homogeneous in virtual multi - dimensional (5-7d) space. The mapping on this virtual space is a 2-layer perceptrons takes into account all predictors.
Hey FBykov, would you mind sharing a little more information in my post here: Care to share general methodologies? ? It would be awesome to hear more about what you did, it sounds like you took a really interesting approach!