Abstract

Nowadays, governments and private sectors request industry software solutions through public tenders that use websites for mass distribution. Not only is demand organized, but a large number of software tenders is produced. This study focuses on the analysis of texts from these documents to characterize them efficiently in order to find a specific solution to the general problem of how to make a bid and how not to make a bid. An automatic classifier is proposed for the public tender process for software based on IEEE standard 830-1998, which categorizes text from a pragmatic point of view. Development phases and classification success rates are shown for each algorithm used in the different experiments. This system may be an alternative for the early analysis of public tenders for software with fuzzy requirements. © 2021, Springer Nature Switzerland AG.