US 12,353,514 B2
Classification of sequential data based on integrated likelihood ratio
Akinori Ebihara, Tokyo (JP)
Assigned to NEC CORPORATION, Tokyo (JP)
Filed by NEC Corporation, Tokyo (JP)
Filed on Aug. 17, 2023, as Appl. No. 18/235,229.
Application 18/235,229 is a continuation of application No. 17/439,009, previously published as PCT/JP2019/012793, filed on Mar. 26, 2019.
Prior Publication US 2023/0401289 A1, Dec. 14, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 18/00 (2023.01); G06F 18/21 (2023.01); G06F 18/2415 (2023.01); G06F 18/2431 (2023.01)
CPC G06F 18/2415 (2023.01) [G06F 18/217 (2023.01); G06F 18/2431 (2023.01)] 4 Claims
OG exemplary drawing
 
1. An information processing device comprising:
at least one memory storing instructions; and
at least one processor configured to execute the instructions to:
(a) acquire one element of sequential data;
(b) acquire a past likelihood ratio calculated in a past for another element acquired before the one element, the likelihood ratio indicating a ratio between a probability that the one element belongs to a first class and a probability that the one element belongs to a second class;
(c) calculate a likelihood ratio indicating likelihood that the element belongs to one of a plurality of classes, based on the one element and the past likelihood ratio;
(d) acquire an integrated likelihood ratio calculated in the past, the integrated likelihood ratio indicating likelihood that the entire sequential data belongs to one of the plurality of classes;
(e) integrate the likelihood ratio calculated for the one element with the integrated likelihood ratio calculated in the past by calculating a logarithm of the likelihood ratio calculated for the one element using the integrated likelihood ratio calculated in the past, and calculate an updated integrated likelihood ratio;
(f) compare the updated integrated likelihood ratio with a threshold, and classify the entire sequential data to a corresponding class when the updated integrated likelihood ratio exceeds the threshold;
(g) repeat processes (a) to (f) if it is when having determined that the entire sequential data cannot be classified into one of the plurality of classes; and
(h) classify the entire sequential data into one of the plurality of classes when having determined that the entire sequential data can be classified into one of the plurality of classes.