I will briefly introduce the technology of "Landscaping".
When analyzing text data, KIBIT identifies parts of speech and extracts words, and calculates the importance of each extracted word in relation to the desired information.When calculating its importance, the presence or absence of relevance is judged using a mechanism called "amount of transmitted information".
Next, the importance of all the words in the text data is totaled and the score is calculated.Sort the scores of that data in descending order.This series of flows is FRONTEO's original algorithm "Landscaping," which is KIBIT's learning and inference method.This algorithm allows KIBIT to learn the tacit knowledge of experts (knowledge based on human experience and judgment).