Timeline for Sequential Form of the Least Squares Estimator for Linear Least Squares Model
Current License: CC BY-SA 4.0
11 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Apr 13, 2019 at 20:26 | comment | added | Royi | This is basically the idea behind Sequential Least Squares I derive in my answer below. | |
| Apr 5, 2019 at 18:58 | comment | added | M529 | @robertbristow-johnson So I guess we all had some fun and OP got an answer. Perfect! :D | |
| Apr 5, 2019 at 18:40 | comment | added | robert bristow-johnson | yeah, and i hadn't put in the time. but it looks like Olli did. | |
| Apr 5, 2019 at 18:02 | comment | added | M529 | @robertbristow-johnson I simply haven't had the time to explicitly formulate it, however I could not resist pointing out the obvious after having checked the link ;) | |
| Apr 5, 2019 at 1:51 | comment | added | robert bristow-johnson | it's definitely doable and i will show you on this answer tonight. did you read about or do you know how line-fitting is done with a Least-Square-Error constraint? | |
| Apr 5, 2019 at 1:43 | comment | added | Izzo | @robertbristow-johnson I believe you are correct. This question was primarily inspired by an "is it possible". I was curious if there was an iterative method to calculate the slope similar to iterative method to calculate a mean value (stackoverflow.com/questions/1930454/…). I'm starting to believe this is not possible. | |
| Apr 5, 2019 at 1:31 | comment | added | robert bristow-johnson | you can let samples that are old enough fall off the edge, so your accumulators don't accumulate too much. | |
| Apr 5, 2019 at 1:29 | comment | added | robert bristow-johnson | are you fixed-point, or is your processor floating-point, Izzo? | |
| Apr 5, 2019 at 1:22 | comment | added | Izzo | I avoided this method due to the fact that I could run into overflow issues. It seems like a running sum could potentially grow too big depending on how many samples are taken thus corrupting the calculation. | |
| Apr 4, 2019 at 19:47 | comment | added | robert bristow-johnson | this was the approach i was gonna make. M529, why don't you slug out an explicit answer and get some booty? i don't need it. | |
| Apr 4, 2019 at 18:10 | history | answered | M529 | CC BY-SA 4.0 |