Top of page

Font Size

Content Area

NTT Press Releases

Fig. 1 What is Jubatus?

Fig. 1 What is Jubatus?

Fig. 2 Flow-type big data × machine learning model

Fig. 2 Flow-type big data × machine learning model

Fig. 3 Why Deep realtime analysis?

Fig. 3 Why Deep realtime analysis?

Jubatus
  Support analysis engine Description Example
Version 1 Classification Classification of given data into categories Identifies spam
Divides tweets into categories
Version 2 Recommendation Recommend data similar to given data Recommends EC site products
Search site-related advertising
Regression Estimate output value of given data Predict power consumption
Predict stock prices
Statistics Summary of statistical data: data frequency, standard deviation, maximum value, minimum value, etc. Monitor sensors
Detect data anomalies
Version 3 Graph mining Extract center and shortest path from given graph structure Analyze social communities
Analyze network structures
Version 4 Outlier detection Detects outliers (anomalies) from given data set Detect fraud, detect faults
Version 5 Clustering Divides given data into sequential clusters Customer clustering

Fig. 4 Functions and Public Sites

Fig. 5 Strengths and roles of three companies

Fig. 5 Strengths and roles of three companies

<< Back to article

  • Search Among NTT Press Releases

     - 

  

  • NTT Press Releases
  • Latest Press Releases
  • Back Number
  • Japanese is here
Footer Area
Copyright © 2013 Nippon Telegraph and Telephone Corporation