NEC develops automatic prediction and analysis tec

  • Detail

NEC has developed automatic prediction and analysis technology

NEC has developed automatic prediction and analysis technology to realize the whole process automation of massive data prediction and analysis process of relational databases widely used in business systems, and the AI technology group NEC the wise (Note 1) has added new members

when analyzing the relational database composed of multiple databases, data analysis experts are required to find out the complex relationships and correlations between databases, and it takes a lot of man hours to learn and adjust the prediction model through machines. However, due to the scarcity of expert resources that can skillfully analyze data, in response to the rapidly growing demand for high data analysis, in order to quickly analyze massive data, we need accurate analysis methods that even non experts can easily grasp

nec automatic prediction and analysis technology can realize the automation of a series of processes from extracting/designing effective data items (eigenvalues) to building the best prediction model. Therefore, even if you do not have a high degree of data analysis ability, you can quickly and accurately make the same or higher level of prediction and analysis as data analysis experts

nec, Sumitomo Mitsui bank and Japan Comprehensive Research Institute jointly verified this technology (Note 2). The results show that this technology can shorten the work that data analysis experts need 2-3 months to complete to 1 day on the premise of maintaining accuracy. In 2017, NEC strives to promote this technology to enterprise users as a service for enterprises to conduct big data analysis internally

nec has been focusing on social solutions business, and hopes to further expand and improve the service business applying this technology, while creating new business opportunities for customers and improving enterprise value through analyzing solutions

nowadays, enterprises all hope to accurately grasp the needs of various customers under the changeable 282 decorative fire retardant coating business environment at the first time, and reflect them in the business measures of the enterprise. By analyzing the huge data accumulated by enterprises to find/create new value, which greatly affects the development of enterprises. On the other hand, the contradiction between the global shortage of data analysis experts and the rapid growth of the demand for massive data analysis makes rapid data analysis an urgent problem to be solved

nec has been developing heterogeneous hybrid learning technology (Note 3) since 2012, which can find the rules mixed in big data. It has been accurately automating the highly manual operation of senior data analysis experts (finding data relevance, from extracting and designing eigenvalues to building analysis models), Developed the pre-processing technology of data analysis, automatic design technology of eigenvalues (released in 2015, the force range of single column universal experimental machine is low, note 4), etc

the automatic prediction and analysis technology released this time is an upgrade of the previous technology, which can fully automatically analyze and predict the massive data of the relational database

advantages of new technology

data analysis process

1 Strengthen the automatic design technology of eigenvalues, automatically find eigenvalues from relational databases

strengthen the automatic design technology of eigenvalues released in 2015, and automatically design eigenvalues from relational databases commonly used in business systems

the important business information of the company is usually divided into multiple tables and stored in the relational database system. AI technology can quickly search and find the permutation and combination values (eigenvalues) of hypothetical effective data from multiple relational databases to be analyzed, in which the generation of essential massive search requests is also automatically completed by the system

therefore, there is no need to rely on the assumptions and plans of analysis experience and related knowledge, and there is no need to operate a large number of operations of generating eigenvalues from data, so as to greatly shorten the analysis time and save analysis manpower. In addition, machine operation is far faster than manual operation, and it can search a large number of assumptions in a short time, so it has the advantages of accurate analysis results and the ability to find new knowledge that people don't notice

2. Through the automatic design technology of prediction model, the best model is automatically selected from multiple models

in addition to the heterogeneous hybrid learning technology based on eigenvalue data, users can also select or combine the best prediction model that can derive the target analysis results from a large number of prediction models based on various mechanical learning techniques such as logical regression and decision tree. In addition, the prediction value calculated by the prediction model will also provide the prediction basis (why such a prediction result is obtained)

because the prediction basis can be mastered, users can make more reasonable plans and judgments according to the situation

this time, an intuitive and operable GUI (graphical user interface) has been developed. The operator only needs to operate according to the prompt screen, and can simply and quickly complete the data eigenvalue exploration and prediction model construction that used to cost a lot of man hours. Therefore, even if you don't have professional data analysis ability, you can make the same or higher level of accurate prediction and analysis as analysts in a short time

this technology helps to analyze the important business information data of the enterprise, formulate strategies based on new potential needs, verify hypotheses, and promote the implementation of policies, so as to contribute to the company's rapid business judgment

nec group is committed to promoting social solutions on a global scale, providing reassuring, safe, efficient and fair social values, integrating advanced ICT technology with knowledge, and contributing to a brighter, richer and more efficient society

(Note 1)

the meaning contained in the sign of NEC the wise

the sign of NEC the wise adopts the simplest triangular cone in the three-dimensional figure, and there is a cube in the center of the triangular cone. The acute triangular cone as the foundation indicates that it is firm and unshakable, while the cube in the center symbolizes AI technology with wisdom. The slanting angle of the trademark reflects the idea of solving all social issues through the coordination of human and human, human and society, human and AI technology, so as to turn instability into stability and create a better society

draft: NEC released the brand name of AI technology NEC the wise

NEC's research on AI http:/the alliance is composed of 8 production enterprises from the province's thermoplastic elastomer material and product industry, 5 universities including Hefei academy of materials science, Chinese Academy of Sciences and Hefei University of technology Composition of scientific research institutes/

(Note 2) draft on December 15, 2016

empirical experiment on automated data analysis using AI technology

(Note 3) heterogeneous hybrid learning technology

June 22, 2012

NEC developed a technology that can automatically discover a variety of rules mixed in big data

June 19, 2014

NEC strengthened the technology that can automatically discover a variety of rules mixed in big data

(Note 4) Automatic design eigenvalue technology

August 18, 2015

NEC developed an automatic design eigenvalue technology that reduces the prediction and analysis time of big data by 1/3

Copyright © 2011 JIN SHI