The results of PLS predicts
Posted: Wed Sep 07, 2022 7:51 am
Hi,
I seek your advice on the results of PLSpredict. If the results of the Q² value in PLSPredict are positive but when comparing with the LM outcomes, majority of the PLS-SEM results have higher prediction errors (e.g., in terms of RMSE or MAE) than the LM. Based on these results, can we still conclude that The PLS-SEM models have better predictive performance than using simply mean prediction but in comparison with LM (linear regression model) outcomes, generally the PLS-SEM models do not improve the the predictive performance of the available indicator data.
Can we just use one of these two benchmarks ( Q2 and RMSE/MAE) or we have to use both to assess the PLS_SEM predictive perfomance?
Appreciate your advice. Thank you.
Best regards,
Lintje
I seek your advice on the results of PLSpredict. If the results of the Q² value in PLSPredict are positive but when comparing with the LM outcomes, majority of the PLS-SEM results have higher prediction errors (e.g., in terms of RMSE or MAE) than the LM. Based on these results, can we still conclude that The PLS-SEM models have better predictive performance than using simply mean prediction but in comparison with LM (linear regression model) outcomes, generally the PLS-SEM models do not improve the the predictive performance of the available indicator data.
Can we just use one of these two benchmarks ( Q2 and RMSE/MAE) or we have to use both to assess the PLS_SEM predictive perfomance?
Appreciate your advice. Thank you.
Best regards,
Lintje