time series forecasting on housing prices.
painpoints of being my first model. not knowing the difference between tabular vs time-series data. for a long time i was cross validating the model just as i would for tabular data, with k-fold sliding window. however, the temporal nature and order of the data is imperative and must be maintained for all train test splits. so i used expanding window k-fold.
updated: 10/17/2024
todo: finish explanation, provide metric images