Time-series forecasting from multiple timeseries
显示 更早的评论
Hello. I am new to machine learning and am unable to find the relevant literature to my problem (probably using wrong keywords). My goal is to predict the most probable label of a series of observations where each observation consists of multiple types and the length of "observation-series" have different lengths.
Currently my dataset looks like the following, Id is the grouping variable of a (consecutive) series of observations, label is the single output that identify.
dataset
Area MajorAxis MinorAxis Label Id
2391 95.42964935 41.92808151 1 035.mpg
2420 96.32376099 42.31892014 1 035.mpg
3248 129.3288422 56.08081818 1 035.mpg
2435 97.15223694 42.97753525 1 035.mpg
2502 99.36376953 42.99647903 1 035.mpg
2461 97.69940948 43.29359055 1 035.mpg
2366 93.56082153 41.85757446 1 035.mpg
2390 95.42179108 41.92354584 1 035.mpg
2390 95.42179108 41.92354584 1 035.mpg
2734 115.0877533 39.17201614 2 036.mpg
3612 149.0444183 53.27929688 2 036.mpg
3652 149.2009125 52.87094498 2 036.mpg
2719 116.4709473 37.98487854 2 036.mpg
2721 117.0030746 37.89451981 2 036.mpg
Is there any built in method in matlab for dealing with multiple input single output timeseries forcasting of this sort?
采纳的回答
Greg Heath
2016-12-15
0 个投票
The MATLAB timeseries equations are valid for vectors. Therefore, multiple inputs and outputs are easily accommodated.
The only drawback is the equations require all inputs to have the same number of delays. In general, this may be very suboptimal.
Hope this helps.
Thank you for formally accepting my answer
Greg
更多回答（0 个）
类别
在 帮助中心 和 File Exchange 中查找有关 Deep Learning Toolbox 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!
Translated by 
