How to obtain a better result of updating the I-O table from the MTT method

Under the assumption of changes in economic structure, Matrix Transformation Technique (MTT) combines matrix transformation and forecasting techniques to update input-output (I–O) tables solving I-O timeliness. This research aims to improve the result of MTT by forecasting only the significant elements instead of forecasting all elements of a transformed-intersection matrix as the original propose. The forecasted elements are considered from their statistically significant trend and their correlation with the technical coefficient. The empirical evidence from updating the Thailand I-O table shows that the performance of MTT with forecasting only significant elements outperforms the result of MTT with forecasting all elements and the well-known method as the RAS approach. The assessment is considered from the updating results of the I-O table, technical coefficient, and multiplier matrices by common indices: Standardized Total Percentage Error (STPE), Theil's U (U), Standardized Weighted Absolute Difference (SWAD), and also Mean Absolute Percentage Error (MAPE).