Honest Error
Each row of X holds the predictors of one observation and y the outcomes, in time order. A model judged on the data it was fitted to looks better than it is. So: fit ordinary least squares with an intercept on the first 80% of the rows (int(0.8 * n) of them), predict the remaining rows, and return the root mean squared error of those predictions.
Example 1
X [[0,0],[1,2],[2,4],[3,1],[4,3],[5,0],[6,2],[7,4],[8,1],[9,3],[10,0],[11,2],[12,4],[13,1],[14,3]]
y [1,13,25,12,24,11,23,35,22,34,21,33,45,32,44]
returns 0.0
The outcomes follow 1 + 2a + 5b exactly, so the held-out rows are predicted perfectly.
Constraints
Xhas 2 columns and 15 to 60 rows
Write it in your browser, against the clock. The model solution, stepped through line by line, comes after.
More statsmodels problems
- Next Three ValuesHard
- Slope Through TimeHard
- The Repeating PartHard
- Which Findings SurviveHard
- How Much Is ExplainedBeginner
- How Sure Is the SlopeBeginner
- Line Through ZeroBeginner
- Line Through the DataBeginner