LLM x Statistical Analysis x Machine Learning

Semantic Data Mining

The essence of LLMs (large-scale language models) can be seen as a technological innovation that makes ‘meaning’ computable. Traditionally, unstructured qualitative information has been considered incompatible with quantitative analysis, but the advent of LLMs has changed this situation significantly. Now, not only can we use LLM to powerfully interpret the results of statistical analysis and machine learning, but it is even possible to vectoriseunstructured qualitative information and analyse it quantitatively through machine learning.

Read more... Tutorials

Build "concepts" from huge pieces of information using Data Science / AI

We provide a completely new business information analysis method for the AI era. Concept Research (=Creative Information Processing) combining LLM(Lerge Language Models), SOM(Self-Organizing Maps) and BBN(Bayesian Belief Network). Applications include: Analysis of Customer Voices at Call Centers, Analysis of Patent Information and Technology Trends, Business Strategy Planning, Case Law Analysis, Accident Analysis, Trouble Information Analysis, Policy Planning, and various Document Management.
Concept Research Method

We have been researching computer-assisted concept creation methods for the past 25 years. About 20 years ago, we realized that SOM and BBN were the most suitable technologies to modernize the KJ method, which was invented in Japan in the 1960s. However, the final piece was difficult to find. The advent of LLM finally gave us that.