Reliability assessment of an intelligent approach to corporate sustainability report analysis

Abstract

This paper describes our efforts in developing intelligent corporate sustainability report analysis software based on machine learning approach to text categorization and illustrates the results of executing it on real-world reports to determine the reliability of applying such approach. The document ultimately aims at proving that given sufficient training and tuning, intelligent report analysis could at last replace manual methods to bring about drastic improvements in efficiency, effectiveness and capacity.