US 12,437,031 B2
Method of hop count matrix recovery based on decision tree classifier
Xingcheng Liu, Guangzhou (CN); Yingying Zhao, Guangzhou (CN); and Yitong Liu, Guangzhou (CN)
Assigned to Sun Yat-sen University, Guangzhou (CN)
Filed by Sun Yat-sen University, Guangzhou (CN)
Filed on Dec. 31, 2021, as Appl. No. 17/566,732.
Claims priority of application No. 202110278022.4 (CN), filed on Mar. 15, 2021.
Prior Publication US 2022/0292318 A1, Sep. 15, 2022
Int. Cl. G06F 18/245 (2023.01); G06F 18/2132 (2023.01); G06F 18/2411 (2023.01); G06F 18/243 (2023.01); H04W 64/00 (2009.01)
CPC G06F 18/245 (2023.01) [G06F 18/21322 (2023.01); G06F 18/2411 (2023.01); G06F 18/24323 (2023.01); H04W 64/006 (2013.01)] 7 Claims
OG exemplary drawing
 
1. A method of hop count matrix recovery based on a decision tree classifier, comprising the following steps:
S1: performing a flooding process to acquire a hop count matrix H˜ with missing entries;
S2: constructing a training sample set according to relationships between a part of observed hop counts in the hop count matrix H˜, and modeling the observed hop counts in the hop count matrix as labels of the training sample set, wherein a maximum hop count represents a number of classes;
S3: training a decision tree classifier according to the training sample set obtained in step S2; and
S4: constructing a feature for an unobserved hop count, to obtain an unknown sample; and inputting the unknown sample to the trained decision tree classifier, to obtain a class of the unknown sample which equals to a missing hop count at a corresponding position in the hop count matrix, so as to recover a complete hop count matrix H;
wherein step S2 specifically comprises:
S201: traversing the hop count matrix H, and performing a next step if a hop count at a certain position is observed; otherwise, traversing a next value in the hop count matrix H;
S202: using node i and node j to represent two nodes between which a hop count is missing, and with respect to all other nodes k, k=1, 2, . . . , n in a network, calculating a minimum hop count sum of a hop count hik from node i to node k and a hop count hkj from node k to node j, as a first feature of a training sample;
S203: calculating an average value of hop counts from neighboring nodes of node i to node j and hop counts from neighboring nodes of node j to node i, as a second feature of the training sample;
S204: using the observed hop count as a class, forming the training sample by using the class together with the first feature and second feature, and adding the training sample to the training sample set; and
S205: traversing the entire hop count matrix H, and obtaining the training sample set after the traversing is finished;
wherein after step S1 and before step S2, if a symmetric position of a missing hop count is observed, the missing hop count is supplemented by using a hop count at the symmetric position;
the hop count at the symmetric position is determined by Normalized root mean square error (NRMSE), which is calculated as follows:

OG Complex Work Unit Math
wherein (xk, ŷk) represents an estimated position of an unknown node k, (xk, yk) represents a real position of the unknown node k, Nu represents the number of unknown nodes, and R is the communication radius of each node.