A Natural Language Processing Approach for the Digitalization of Roaming Agreement

Roaming is the capability of a subscriber to have persistence of service in a visited mobile network managed by a different operator. To achieve this persistence of connectivity, it is necessary for operators to agree on technical, commercial and legal aspects in what is known as a Roaming Agreement(RA). Currently, RA negotiation uses asynchronous flow such as email or even regular mail. The main phase in building the RA is known as the drafting phase, which is a time-consuming process with limited assurance and transparency on handling the drafting events. To address the lack of transparency and reduce the drafting time, this paper proposes a framework that automates the process of drafting and negotiation of the RA and enables better accessibility. The proposed framework exploits a Natural Language Processing (NLP) engine as a starting point for the digitalization of the negotiation process towards transparent drafting of RAs. Extensive experiments on real RAs show the superiority of our proposed framework in drafting the RAs against the traditional drafting and negotiation process with prediction accuracy over 80.0%.