Search code examples
time-seriessvmpredictioncontinuous

Is possible to run support vector machine using following data to predict survival?

Our data contain three types of independent variables.

Age: categorical variable

Sex: continuous variable

Lab1wk / Lab2wk / Lab4wk / Lab6wk: time series data of a specific laboratory value

------------Age-Sex-Lab1wk-Lab2wk-Lab4wk-Lab6wk---Survival
Case1--------31--1--------23-------21--------2--------31----------Yes

Case2---------2--2--------31-------31-------39--------92----------No
Case3---------9--1--------21--------0-------31--------34----------Yes
.....
Case3092---42--1--------93-------99-------32---------0----------Yes

To predict the survival I am trying to run support vector machine using python.

And I want to include all lab values (Lab1wk / Lab2wk / Lab4wk / Lab6wk) in the model, because I think the unknown pattern of the lab values strongly affects the survival.

With approximately 3000 cases, is it reasonable to run support vector machine for the prediction of survival?

Or should I run another statistical method?

Thank you very much, in advance.

Solution

  • With the help of the information within this site, I solved this problem by ensemble model using deep neural network and recurrent neural network. Thank you~!