? Features of Extended Bayesian Network Services Modellize, Inc.

Features of Extended Bayesian Network Services

A Bayesian network is:

A method of representing indefinite and continuously changing variables that cannot be accurately expressed with numerical formulas, such as customersf purchasing behaviors. It uses nodes with probabilities and dependencies between nodes to generate a model in the form of a probability network.

Features and Services

Modellize developed its own behavior prediction engine and provides services for modeling and predicting customers behavior, based on the following achievements after 10 years of research at AIST:

- Technology to automate and simplify the development of Bayesian networks
- Technology to predict behaviors using developed models
- High-speed processing (algorithms, grid computing)

Opportunities for Your Company

Customer purchasing behavior models developed can be commonly used and applied in various areas as follows.

Prediction of target customers and products:
recommendation of products and services via PCs or cell phones, on-line advertisements, sales person support

Trouble shooting, problem solving, new product planning, customer evaluation and selection, various direct marketing activities

Features of Extended Bayesian Network Services

The customer models we develop can be commonly applied to various consumer services.

Analysis based on Bayesian Approach Customers Sales staff Traditional analysis - Customer data - Product data - Purchase record data Existing database  Customer analysis knowledge Product sales know-how  Questionnaire Interview Growth and update of customer models using the latest data Bayesian network  Customer models Customer models can be commonly and comprehensively used for various purposes Behavior prediction engine Sales staff support (recommendations assistance) Online recommendations (PCs, cell phones) Direct marketing (various targeting) Planning for new products and supplies Predicting the effect of advertisements and promotional activities Evaluating customers and promoting membership Trouble shooting and presenting solutions Sales and Profit Maximization

Flow of Modellize Service

Step 1 Consulting ・Identify business issues ・Select pilot services ・Establish service objectives ・Do research on existing data DataAnalysis ・Establish analytical targets Perform interviews and questionnaire ・Preprocessing of data ・Develop Bayesian network models ・Perform at-desk simulation and model enhancement ・Prepare analysis reports Step 2 Validation Testing ・Define service specifications ・Prepare for provision of service results ・Perform model update and analysis based on latest data ・Conduct validation field testing ・Prepare reports on test results Step 3 Service AP Development ・Define final service specifications ・Develop service AP development plans ・Design additional functions to be developed for service AP ・Develop Service AP ・Conduct quality assurance testing Service AP Validation Testing ・Define target objectives ・Update final model ・Conduct final validation testing ・Achieve goals through model enhancements ・Prepare report on test results Step 4 Operation Your Company ・Provide services to customers Operate service sites, etc. Modellize, Inc. Develop initial models, Continue updates ・Perform analysis and provide recommendations using the behavior-prediction server