Unlock instant, AI-driven research and patent intelligence for your innovation.

Apparatus and program for detecting abnormal behavior

a program and abnormal behavior technology, applied in the field of probabilistic distribution estimation apparatus, abnormal behavior detection apparatus, abnormal behavior detection apparatus, etc., can solve the problems of inefficient calculation method according to k. yamanishi and j. takeuchi, inability to adapt to a variation of an information source for generating data in the prior arts, and insufficient scalability

Inactive Publication Date: 2007-11-08
NEC CORP
View PDF2 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Benefits of technology

[0023] It is a third object of this invention to provide a probabilistic distribution estimation apparatus and method and an abnormal behavior detection apparatus and method which are capable, in a problem dealing with behavior data into which a plurality of programs or data of a plurality of users are mixed, not only of detecting a variation of individual behavior but also of detecting abnormal behavior data at a high precision when the behavior data is few.

Problems solved by technology

However, there are problems in the above-mentioned prior arts as follows.
A first problem is no adaptability for a variation of an information source for generating data in the prior arts.
A second problem is no sufficient scalability.
GB 2361336 A and the method according to K. Yamanishi and J. Takeuchi are inefficient on calculation as well as necessary of the large scale of memory capacity because a leaning algorithm uses all past data in there methods.
A third problem is no robustness for noises.
A fourth problem is that abnormal behavior enable to detect is restricted.
A fifth problem is that precision of detection is bad in a case of few data.
A sixth problem is that data of analysis target is restricted.
Likewise, the system according to K. Yamanishi and J. Takeuchi cannot detect abnormality in a pattern of behavior data although the system can detect outlier or a variation point in the discrete data or the continuous data one-point by one-point.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Apparatus and program for detecting abnormal behavior
  • Apparatus and program for detecting abnormal behavior
  • Apparatus and program for detecting abnormal behavior

Examples

Experimental program
Comparison scheme
Effect test

second embodiment

[0192] Referring to FIG. 3, the description will proceed to a probabilistic distribution estimation apparatus 4 according to a second embodiment of this invention. The probabilistic distribution estimation apparatus 4 is similar in structure and operation to the probabilistic distribution estimation apparatus 2 illustrated in FIG. 1 except that the probabilistic distribution estimation apparatus 4 further comprises a session unit 41. Inasmuch as operations of the certainty calculation unit 21, the parameter storage unit 22, the parameter renewal unit 23, and the parameter output unit 24 are similar to those of the units 21-24 in the probabilistic distribution estimation apparatus 2, the operations thereof are omitted.

[0193] When date obtained from the input unit 1 has no vector format indicative of behavior, the session unit 41 carries out session for converting the data into vector data. In other words, the session unit 41 processes or converts the input data into the string of ve...

third embodiment

[0197] To express the stochastic model having the above-mentioned hidden variable is sufficient to appoint values of the parameters calculated by the probabilistic distribution estimation apparatus 2 or the probabilistic distribution estimation apparatus 4. Accordingly, an abnormal behavior detection apparatus receives the values of the parameters from the parameter output unit 24 of the above-mentioned probabilistic distribution estimation apparatus 2 or the above-mentioned probabilistic distribution estimation apparatus 4 to calculate an abnormal behavior-like of the input data using the values of the parameters.

[0198]FIG. 5 is a block diagram showing an abnormal behavior detection apparatus according to a third embodiment of this invention. The abnormal behavior detection apparatus comprises the input unit 1 for inputting data, the output unit 3 for outputting a state estimated score, either the probabilistic distribution estimation apparatus 2 illustrated in FIG. 1 or the proba...

fourth embodiment

[0203] Referring to FIG. 7, the description will proceed to an abnormal behavior detection apparatus according to a fourth embodiment of this invention. The abnormal behavior detection apparatus according to the fourth embodiment comprises the input unit 1 for inputting data, the output unit 3 for outputting an abnormal behavior degree, either the probabilistic distribution estimation apparatus 2 illustrated in FIG. 1 or the probabilistic distribution estimation apparatus 4 illustrated in FIG. 3, and an abnormality detection unit 6 for calculating an abnormal behavior degree of the input data. In other words, the abnormality detection unit 6 calculates an information amount of the new read data by using the parameters of the probabilistic distribution estimated by the probabilistic estimation apparatus to produce the abnormal behavior degree of the new read data.

[0204] The abnormality detection unit 6 comprises an abnormal behavior degree calculation unit 61 which calculates and ou...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

Supplied with a string of vector data as input data, a probabilistic distribution estimation apparatus estimates, by using a stochastic model having hidden variables, a probabilistic distribution in which each data occurs by successively reading the train of vector data. Specifically, the probabilistic distribution estimation apparatus reads values of parameters of the stochastic model having the hidden variables for a value of the input data, calculates, by using the stochastic model, a certainty in which the input data occurs, renews the parameters in response to new read data with past data forgotten, and produce several parameter's values. By using the parameter's values received from the probabilistic distribution estimation apparatus, an abnormality detection unit calculates an information amount of data as an abnormal behavior degree to produce the abnormal behavior degree.

Description

CROSS-REFERENCE TO RELATED PATENT APPLICATIONS [0001] This application is a division of application Ser. No. 10 / 778,178, filed Feb. 17, 2004, now pending, and related to four concurrently filed applications entitled ABNORMAL BEHAVIOR DETECTION APPARATUS (Attorney Docket No. 016778-0518), METHOD OF DETECTING ABNORMAL BEHAVIOR (Attorney Docket No. 016778-0519), APPARATUS AND METHOD OF DETECTING ABNORMAL BEHAVIOR (Attorney Docket No. 016778-0520), and APPARATUS, METHOD AND PROGRAM FOR DETECTING ABNORMAL BEHAVIOR (Attorney Docket No. 016778-0521), and based on Japanese Patent Application No. 2003-40347, filed Feb. 18, 2003, and Japanese Patent Application No. 2003-171481, filed Jun. 17, 2003, by Yuko Matsunaga and Kenji Yamanishi, which are incorporated herein by reference in their entirety. This application claims only subject matter disclosed in the parent application and therefore presents no new matter.BACKGROUND OF THE INVENTION [0002] This invention relates to a probabilistic dist...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(United States)
IPC IPC(8): G06F17/10G06F7/00G06F17/18G06K9/62G06N3/00G10L15/14
CPCG06K9/6297G06F17/18G06F18/295
Inventor MATSUNAGA, YUKOYAMANISHI, KENJI
Owner NEC CORP