BAI4SOW: Soft Computing for supporting Business Intelligence in Social Workflows

Social network workflows (SOW) coordinate the activities performed by a set of users that individually or cooperatively try to reach a particular objective. SOWs are non-structured workflows that involve a great number of users performing a wide range of activities throughout a time period, typically using a low amount of computational resources. Some examples of processes modelled by this kind of workflows are Marketing Campaigns -aimed to encourage clients to consume a particular product or service-, Employee Training or Gamification.
The SOCOBAI project (Soft computing as a tool for Business Artificial Intelligence), combines different Process Mining techniques that allow: (i) to automatically discover the real workflow executed by the users; (ii) to extract relevant information of this workflow, as the frequent patterns of the activities, which reveal valuable information about user behavior; (iii) to automatically hierarchize non-structured workflows with the aim of enhancing the visualization of processes by reducing their complexity in different abstraction levels; and (iv) the linguistic description of those workflows, through which the system is capable of describing the most relevant issues of the workflows in Natural Language, generating supplementary ‘friendly’ data to complement the visual information provided by workflows.
SOCOBAI is part of the BAI4SOW project: Soft Computing for the support of Business Intelligence in Social Workflows. involving the CiTIUS ( Project Leader), the University of Alcalá and the University of Zaragoza.


The main goal of the BAI4SOW project is the development of intelligent techniques for the automatic extraction and analysis of user’s behavior in social workflows. Moreover, the specific objectives of the project are the following:

  1. To develop Process Mining algorithms in order to automatically discover precise and simple models aimed to describe the workflow of activities performed by users in social workflows, as well as to predict the duration of the different activities involved in complex Marketing Campaigns.
  2. To define a model and to develop a computer application for the automatic generation of text reports (linguistic reporting) capable of describing the situation and evolution of the processes followed by users in social workflows, and so providing the relevant information about these processes to the main actors.
  3. To develop automatic learning algorithms for the recommendation and automatic extraction of the user profiles in social workflows