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

Support vector machine classifier IP (internet protocol) core

A support vector machine and support vector technology, applied in the field of microelectronics, can solve the problems of limited quantity and unsatisfactory performance, and achieve the effects of fast transmission speed, improved flexibility, and reduced training costs

Inactive Publication Date: 2015-05-06
SHANGHAI UNIV
View PDF4 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the actual sample size is often limited, so its performance may not be satisfactory in practical applications.

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
  • Support vector machine classifier IP (internet protocol) core
  • Support vector machine classifier IP (internet protocol) core
  • Support vector machine classifier IP (internet protocol) core

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0032] like figure 1Shown, the support vector machine classifier IP core of the present invention comprises training module (1) and classification module (2); The input terminal of training module (1) is connected with external data line, receives training sample data, and the training module (1) The output end is connected with the first input end of the classification module (2), and the second input end of the classification module (2) is connected with the external data line to receive the sample data to be classified, and the output end of the classification module (2) is a support vector machine classifier Classification result output terminal of IP core. Training module (1) After receiving the training sample data through the data line, the sample data is first split into two parts, the sample label and the sample data, and then the inner product between the two samples is calculated, and finally according to the inner product and the sample The label calculates th...

Embodiment 2

[0034] This embodiment is basically the same as Embodiment 1, and the special features are as follows:

[0035] see Figure 2~Figure 5 , the training module (1) includes a sample split unit (1-1), a sample label memory (1-4), a sample data memory (1-2), an inner product calculation unit (1-3), an inner product memory (1-5), classifier parameter calculation unit (1-6); the two output terminals of the sample splitting unit (1-1) are respectively connected with the sample label storage (1-4) and the sample data storage (1-2 ), the first output terminals of the two memories (1-2, 1-4) are connected with the two input terminals of the inner product calculation unit (1-3), and the output terminals of the inner product calculation unit (1-3) are connected with the inner product calculation unit (1-3) The inner product memory (1-5) is connected, and the second output end of the sample label memory (1-4) and the sample data memory (1-2) and the output end of the inner produ...

Embodiment 3

[0037] This embodiment is basically the same as Embodiment 2, and the special features are as follows:

[0038] see Figure 6~Figure 8 , the classification module, including lagrange multiplier memory (2-1), classification threshold value memory (2-2), support vector memory (2-3), inner product calculation unit (2-4), the second inner product memory (2-5) and classification unit (2-6); Three output terminals of upper level training module (1) are respectively connected with lagrange multiplier memory (2-1), classification threshold value memory (2-2), support vector The input end connected to the memory (2-3) is connected, the output end of the support vector memory (2-3) is connected to the first input end of the inner product calculation unit (2-4), and the inner product calculation unit (2-4) The second input terminal is connected with the external data line, receives the sample data to be classified, and calculates the inner product of the sample to be classified ...

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

The invention relates to a support vector machine classifier IP (internet protocol) core which comprises a training module and a classifying module. The training module acquires data such as training samples through a data bus from the outside, the data of the training samples are split into sample data and sample tags by a sample splitting unit, the sample data and the sample tags are stored into a designated sample data storage and a designated sample tag storage respectively, the corresponding data of designated samples are respectively read by an inner product calculation unit from the sample data storage and the sample tag storage, obtained calculation results are stored in a designated inner product storage, needed relevant data are read by a lagrange multiplier and a classifying threshold calculation unit from the sample data storage, the sample tag storage and the inner product storage, the lagrange multiplier and a classifying threshold are calculated, the classifying module is connected with the training module to obtain the lagrange multiplier, the classifying threshold and support vectors needed by classification, a classifying unit acquires sample data to be classified through the data bus from the outside, the needed relevant data are read from the storages, and the samples are classified according to classifying functions.

Description

technical field [0001] The invention belongs to the field of microelectronics, and mainly relates to a support vector machine (Support Vector Machine, SVM) classifier IP core, specifically a classifier based on statistical learning theory and obtained through machine learning. With strong classification ability, it can be applied to face detection, text recognition, biological sequence analysis, etc. Background technique [0002] Data-based machine learning is an important aspect of modern intelligent technology. It seeks laws by observing data (samples), and uses these laws to predict future data or unobservable data, including machine learning methods such as pattern recognition and neural networks. , all so. [0003] One of the common theoretical foundations of such machine learning methods is statistics, an asymptotic theory as the number of samples tends to infinity. However, the actual sample size is often limited, so its performance may not be satisfactory in prac...

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(China)
IPC IPC(8): G06K9/62
Inventor 徐美华沈华明沈东阳冉峰
Owner SHANGHAI UNIV