We are working on device failure detection using syslogs and trying to use HSMM Model for the same .We have trained HSMM models one with Error sequences and another with Errorfree sequences. Now with the saved models we are trying to test a new set of observations to get the loglikelihood for it being a Error or Errorfree sequence. Below are the loglikelihoods for both Error and Errorfree data being trained with both Error and Errorfree models. Please help in inferring the below loglikelihoods values for 3 sample sequences of Error and Errorfree data. We also want to know if higher loglikelihood is better for a particular model.
error_e_norm represents error data trained with error model
error_ef_norm represents error data trained with errorfree model
errorfree_e_norm represents errorfree data trained with error model
errorfree_ef_norm represents errorfree data trained with errorfree model
We are working on device failure detection using syslogs and trying to use HSMM Model for the same .We have trained HSMM models one with Error sequences and another with Errorfree sequences. Now with the saved models we are trying to test a new set of observations to get the loglikelihood for it being a Error or Errorfree sequence. Below are the loglikelihoods for both Error and Errorfree data being trained with both Error and Errorfree models. Please help in inferring the below loglikelihoods values for 3 sample sequences of Error and Errorfree data. We also want to know if higher loglikelihood is better for a particular model. error_e_norm represents error data trained with error model error_ef_norm represents error data trained with errorfree model errorfree_e_norm represents errorfree data trained with error model errorfree_ef_norm represents errorfree data trained with errorfree model
error_e_norm , error_ef_norm , errorfree_e_norm , errorfree_ef_norm -14953.29 , 301.81 , -14663.67 , 301.26 -15640.07 , 301.84 , -15391.39 , 295.78 -15686.31 , 301.85 , -16370.14 , 293.46