Tube Loss: A Novel Loss Function for Prediction Interval Estimation
Pritam Anand p r i t a m _ a n a n d @ d a u . a c . i n pritam\[email protected], Tathagata Bandyopadhyay t a t h a g a t a _ b @ d a u . a c . i n tathagata\[email protected], Suresh Chandra s u r e s h i i t d e l · 2024
Read Tube Loss: A Novel Loss Function for Prediction Interval Estimation by Pritam Anand p r i t a m _ a n a n d @ d a u . a c . i n pritam\[email protected], Tathagata Bandyopadhyay t a t h a g a t a _ b @ d a u . a c . i n tathagata\[email protected], Suresh Chandra s u r e s h i i t d e l (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-06853.txt.
Description
Pritam Anand, Tathagata Bandyopadhyay, and Suresh Chandra design a loss function that draws better prediction intervals. Their paper appeared in two thousand twenty-four. A prediction interval gives a range likely to hold a future value. Good intervals hit their promised coverage without sprawling wide. Existing methods trade those goals awkwardly. Skewed data makes the trade worse.
Their Tube Loss trains both interval bounds at once. Theory shows the intervals reach the chosen confidence level as data grows. A tuning knob slides the interval up or down. Sliding lets users center the range where outcomes cluster densely. That sharpens width, especially under skew. Forecasts in finance, weather, and demand planning all use such intervals.
Catalog record
| CREATOR | Pritam Anand p r i t a m _ a n a n d @ d a u . a c . i n pritam\[email protected], Tathagata Bandyopadhyay t a t h a g a t a _ b @ d a u . a c . i n tathagata\[email protected], Suresh Chandra s u r e s h i i t d e l |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 90 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-06853