What applications for the analysis of feelings?
The combination of a volume of information exponential growth in online and exploration tools, analysis and processing of mass data, structured or not, pushes the emergence of a new class of solutions of "feelings analysis. Seth Grimes, a specialist subject, gives us, in an article InformationWeek, some keys to understanding the uses possible.
start by illuminating what the "sentiment analysis", in proposing this definition (adapted from that proposed on the site of " Sentiment Analysis Symposium"): the automated study feelings and opinions expressed in unstructured documents . To achieve this goal, several techniques are put to: exploration and information retrieval, semantic analysis and filtering concepts to identify and filter the relevant content (or remove the "noise"), categorization and consolidation of results ...
Intuitively, one understands the importance of this discipline, which promises to extract the "essence" of the web today, with its countless blogs, social networks, forums ... where surfers from around the world express their views and influence their peers. When current study skills (particularly surveys or focus groups ), timely, require significant resources, "analysis of feelings" to measure in quasi-real time and effortlessly opinions floating around, a sector a product, brand ... The
uses the most obvious interest to marketing departments naturally or those (if any) that focus on "customer experience" to listen to the market (think that all actors concerned to an industry, competition ...) or listen to customers (more focused on the perception of the company and its products or services). But the "analysis of feelings" may also be involved in the measurement of employee satisfaction (not least by automating the study of responses to internal investigations) or become a tool for decision support on capital markets (where the "feeling" naturally plays an important role in guiding the way).
Let us not however believe that the "analysis of feelings" is a universal solution! It must be used wisely as part of a set tools and techniques. For example, detecting a negative trend should be depth, in consultation with specialists who will highlight the causes of this development and will help make an effective response.
conclude this overview of the topic with some examples of market offerings : French Digimind (sector analysis, competitive), Clarabridge (customer focus), RavenPack (news analysis for financial professionals). And do not forget that there are dozens more ...
0 comments:
Post a Comment