Machine learning of SVM classification utilizing complete binary tree structure for PAM-4/8 optical interconnection

2017 ◽  
Author(s):  
Guoyao Chen ◽  
Lin Sun ◽  
Ke Xu ◽  
Jiangbing Du ◽  
Zuyuan He
Keyword(s):  
Machine Learning ◽  
Binary Tree ◽  
Tree Structure ◽  
Optical Interconnection ◽  
Complete Binary Tree ◽  
Svm Classification ◽  
Binary Tree Structure

Heap Adjustment Algorithms Based on Complete Binary Tree Structure

2013 ◽  
Vol 321-324 ◽  
pp. 2122-2127
Author(s):  
Yuan Hu
Keyword(s):  
Binary Tree ◽  
Priority Queue ◽  
Tree Structure ◽  
Complete Binary Tree ◽  
Data Set ◽  
Tree Node ◽  
Linear List ◽  
Data Volume ◽  
Binary Tree Structure ◽  
Node Position

In many applications, it is necessary to adjust the node position of complete binary tree, making the data set have the nature of heap. Because when binary tree node data volume is huge, it will be needed apparently for heap adjustment efficiency of repair algorithms which is based on heap linear list to be improved. Now an improved algorithm is proposed by experiments, which utilizes characteristics of priority queue (delete the earliest data) and stack (delete the latest data), and conducts stack pushing on binary tree node, then accesses node successively to call glide adjustment algorithm to improve the adjustment efficiency.

A SIMPLE ARRAY VERSION OF M-HEAP

2014 ◽  
Vol 25 (01) ◽  
pp. 67-88
Author(s):  
HAEJAE JUNG
Keyword(s):  
Open Problem ◽  
Binary Tree ◽  
Tree Structure ◽  
Complete Binary Tree ◽  
Indexing Scheme ◽  
Binary Tree Structure

Both the post-order heap and the M-heap have a full binary tree structure and have constant amortized insertion and O(logn) deletion time complexities. This paper proposes a simple array version of the M-heap, called AM-heap. The AM-heap has a complete binary tree structure and its array indexing scheme is the same as the simple indexing scheme of the conventional binary heap. An insertion on an AM-heap takes constant amortized time and a deletion takes O(logn) time where n is the number of elements in an AM-heap. The AM-heap resolves the open problem that is to design an array version of the M-heap. Also, it is simpler than the post-order heap to implement and debug.

Parametric Design of Parts in an Integrated Simulation System Based on Asynchronous Mode of Pro/TOOLKIT

2012 ◽  
Vol 479-481 ◽  
pp. 1403-1408
Author(s):  
Gang Lian Zhao ◽  
Yi Jiang ◽  
Yu Jun Chen ◽  
Yan Li Ma
Keyword(s):  
Binary Tree ◽  
Parametric Design ◽  
Design Procedure ◽  
Simulation System ◽  
Tree Structure ◽  
Asynchronous Mode ◽  
Engineering Drawing ◽  
Integrated Simulation ◽  
Binary Tree Structure ◽  
Different Parts

Based on software Pro/ENGINEER and Visual C++ 2005，sub-module of parametric design of assembly with wide universality was done by using Pro/TOOLKIT, and the design procedure was introduced in details. Assembly relation of sub-components is transformed into binary tree structure to store and search parts, and the assembly relation is displayed by CTreeCtrl control. The corresponding parts can be quickly found in the binary tree. Engineering drawing was automatically generated and displayed by ProductView after loading a part, and in this way dimensions of different parts can be modified according to engineering drawing in asynchronous mode. The sub-module can meet the needs of parametric design of parts in the integrated simulation system.

Piecewise linear classifiers using binary tree structure and genetic algorithm

1996 ◽  
Vol 29 (11) ◽  
pp. 1905-1917 ◽  
Author(s):  
Bing-Bing Chai ◽  
Tong Huang ◽  
Xinhua Zhuang ◽  
Yunxin Zhao ◽  
Jack Sklansky
Keyword(s):  
Genetic Algorithm ◽  
Binary Tree ◽  
Piecewise Linear ◽  
Tree Structure ◽  
Linear Classifiers ◽  
Binary Tree Structure

Color Quantization Based on Hierarchical Frequency Sensitive Competitive Learning

2010 ◽  
Vol 14 (4) ◽  
pp. 375-381
Author(s):  
Jun Zhang ◽  
◽  
Jinglu Hu
Keyword(s):  
Binary Tree ◽  
Competitive Learning ◽  
Tree Structure ◽  
Experimental Results ◽  
Color Quantization ◽  
Adaptive Procedure ◽  
Root Node ◽  
Binary Tree Structure

In this paper, we propose a Hierarchical Frequency Sensitive Competitive Learning (HFSCL) method to achieve Color Quantization (CQ). In HFSCL, the appropriate number of quantized colors and the palette can be obtained by an adaptive procedure following a binary tree structure with nodes and layers. Starting from the root node that contains all colors in an image until all nodes are examined by split conditions, a binary tree will be generated. In each node of the tree, a Frequency Sensitive Competitive Learning (FSCL) network is used to achieve two-way division. To avoid over-split, merging condition is defined to merge the clusters that are close enough to each other at each layer. Experimental results show that the proposed HFSCL has desired ability for CQ.

Sign in / Sign up

Export Citation Format

Share Document