Fujitsu Releases QRMining Risk Analysis and Forecasting Software

Tokyo, December 16, 2011

Fujitsu Limited and Fujitsu Chubu Systems Limited today announced the release of QRMining, a software package, primarily for manufacturers of consumer products such as automobiles and electronics, which analyzes and predicts in advance major product issues.

By enabling the integrated monitoring and analysis of various data sources—including information from consumers that has been made public by the government, consumer complaints from a company's call center, and Internet reviews from social media as Twitter, blog sites, and other sources—the new software makes it possible to detect early warning signs of major problems with a company's commercially available products and quickly implement countermeasures.

The software employs text mining technology⁽¹⁾ from Fujitsu Laboratories Limited that extracts details of complaints from text-based malfunction-related data. It also utilizes complex sequential analysis technology that performs multifaceted analysis of a wide variety of data, thereby enabling the discovery of hidden causal relationships and the detection of early signs of abnormality.

During the development of this software, Fujitsu collaborated with customers from the automobile industry to confirm the results of the technology through joint testing.

Background

In recent years, there has been a growing need in the automobile, electronics, and other manufacturing industries to prevent the occurrence and spread of problems and accidents by quickly detecting malfunctions with products available on the market and rapidly implementing necessary countermeasures. Moreover, it is not sufficient to solely keep track of consumer complaints. Companies are also facing an increasing demand for integrated monitoring and analysis of a wide range of information related to product malfunctions.

It is these factors that prompted Fujitsu and Fujitsu Chubu Systems to offer QRMining, a software package that preemptively analyzes and predicts major issues with a company's products.

Technology Employed in QRMining

The new software leverages complex sequential analysis technology that has been developed by Fujitsu Laboratories and performs multifaceted analysis of a wide variety of data in order to discover hidden causal relationships and detect early signs of abnormal circumstances. The technology performs the following three steps in performing a comprehensive analysis of the relationships between complaint information and event information (spread of harmful rumors, mandatory recalls, stock price fluctuations, etc.), thereby enabling highly-accurate sign detection and prediction.

1. Utilizing text mining technology to extract the content of complaints from the descriptive text portion of claims data.
2. Leveraging machine learning technology to compile and simulate the features and changes in complaint information that serve as warning signs of events occurring.
3. Predicting future events by running recent complaints through a forecast model that has automatically learned from past events.

Product Features

1. Employs proprietary technology to monitor and analyze market sentiment of a company's products with high precision

The software performs integrated monitoring and analysis of the market sentiment of a company's products using a variety of data, including information from consumers that has been made public by the government, consumer complaints from a company's call center, and Internet reviews from social media such as Twitter, blog sites, and other sources. This, in turn, allows hidden causal relationships to be revealed and malfunctions to be detected. By leveraging complex sequential analysis technology from Fujitsu Laboratories, the software can detect and analyze warning signs with a high degree of precision.

2. Forecasts the near-future reactions of consumers using consumer behavior modeling

The software develops a model based on the behavior and reactions of consumers towards previous product problems. It then utilizes this model to forecast changes in and the course of future consumer reactions.

3. Forecasts risk from three perspectives: quality of the product, quality of the market, and quality of the company

The software performs an analysis from three perspectives: 1) "Product quality," which predicts risks arising from product malfunctions; 2) "Market quality," which predicts risks arising from consumer overreaction; and 3) "Company quality," which predicts important factors that have the potential to tarnish a company's brand.

Pricing and Availability

Pricing and Availability

Product Name	Monthly software usage fee	Availability
QRMining V1 AS Type, including 10 basic user licenses	1.22 million JPY/month	Late December 2011
QRMining V1 AS Type, per additional user license	32,000 JPY/month	Late December 2011