Opinion Analysis in Interactions

2Analyzing User Opinions in Human–human Interactions

In this chapter, we shall consider scientific problems in the context of customer relationships management, based on studies carried out at EDF Lab. An overview of these studies is shown in Figure 2.1.

Figure 2.1. Opinion analysis and extraction of business concepts from client data from a range of sources (web, call center, etc.)

First, we shall consider our first research question concerning the relevance of different theoretical models in constructing a computational model for different applications (social network analysis, customer relationships management, recommendations) (Q1, see section I.1.3), describing the different business concepts involved, notably the concept of satisfaction as a more limited form of the opinion concept.

However, most of this chapter is given over to our second research question, concerning the capacity of different methods to create generic and relevant computational models as a function of the application and data involved (Q2a and b, see section I.2.6). We shall present different computational models constructed from human–human oral and written interaction corpora, within the context of customer relationships management and online trend analysis.

Using the oral interaction corpora (in this case, from call centers and satisfaction surveys) and written interaction corpora (forums) described ...