How to Incorporate Upward Trend into XGBoost Time Series Forecasting

External question — Data Science Stack Exchange Author: C Dixon Original post: https://datascience.stackexchange.com/questions/112086 License: CC BY-SA 4.0 — https://creativecommons.org/licenses/by-sa/4.0/ Adaptation: HTML converted to plain text; contact email addresses removed. I'm working with an XGBoost XGBRegressor model right now, attempting to utilize it to predict time-series forecasted data. My dataset is not publicly available, so I will use general terms to describe it: It is working to predict at most 2.5 years out, working on 4.5 years of data to predict this. The dataset's target values go up and down seasonally, yet has a constant upward trend throughout the entire dataset (think the stock market, or average earth temperatures). My XGRegressor model is able to capture these intra-yearly trends (the yearly up's and down's) quite well, however is unable to notice the overall upward trend. If a line of best fit were to be applied to my current predicted data, it would be horizontal, and I need it to have a positive slope. I've heard rumors of people able to incorporate linear regression trends into their XGBRegressors, or using two models at once, however I can't find anyone that knows this. However, just a simple line-of best fit incorporation would work just fine for now. Thanks!