A presentation regarding applied techniques and tools for advanced analysis targeted at improving businesses but also with the objective of preventing misuse and unintended behaviours was held yesterday at the Learn@CROZ education centre.
During the presentation, experts from CROZ and IBM explained how predictive analytics and early fraud detection can be key to a business’ success, using many examples and with instruction directed at the banking and insurance industries, as well as toward state institutions and marketing firms.
Right from the start, Luka Stepinac, employing the SentiMenter tool for advanced analysis of sentiment, showed us everything that can be deduced from the mass of data that surrounds us at all times.
A guest from IBM, our colleague Klemen Košelj, stated in his first lecture the importance of predictive analysis in daily business activities through samples of world-wide practices.
Darko Benšić presented the Watson Explorer tool, who’s function is real-time research, analysis and interpretation of data and to display it’s computations singularly. This tool is exceptional of a 360 Customer perspective since it offers the ability to answer in real-time why and how something has occured using advanced cognitive services with neural networks, tekst and market opinion analysis. When Watson Explorer is updated with predictive analysis models, it interactively predicts what may happen, for instance, which product will the user buy next.
The very purpose of predictive analytics with emphasis on why it is employed, i.e. what it actually offers us was presented by Vladimir Mihaljević. Through examples of its practical applications with segmentation and classification models, he showed everything that was necessary to secure for adequate and successful implementation of predictive analysis as well as its eventual influence on the complete optimization of business processes. To that end he suggested the IBM SPSS Modeler predictive analytics tool, which like other IBM products, is configured for simple real-time deployment (ADM, CADS).
Entertaining and yet useful interaction finished off the presentation, where we analyzed some actual cases mentioned during the event, the with the help of SentiMenter.
