Competitor Recognition Based on Improved SVM Algorithm in E-commerce
Journal Title: 河南科技大学学报（自然科学版） - Year 2016, Vol 37, Issue 1
Abstract
To solve the low accuracy of identification of potential competitors in electronic commerce,an improved( support vector machine,SVM) algorithm was presented to optimize the recognition. By combining the Mercer theorem with the factor of margin,the kernel function was modified step by step in order to improve the generalization ability of classifier. Taking Listing information extracted from ebay as the object of research,the kernel function was transformed through conformal mapping based on the Mercer theorem firstly. Then based on the Riemannian geometry theory,the kernel function was improved by integrating the interval factors to form a new classifier. Finally the parameter was chosen automatically to determine the model and test the data. Compared with vector similarity algorithm and the original SVM algorithm,the accuracy of improved algorithm to identify competitors is increased by 10. 2% and 3. 8% respectively.
Authors and Affiliations
Ruili SUN, Shengshuang CHEN, Shijun LI
Keywords
Related Articles
- EP ID EP459951
- DOI 10.15926/j.cnki.issn1672-6871.2016.01.010
- Views 51
- Downloads 0
Shear Capacity Prediction of Reactive Powder Concrete Beams Based on Neural Network
In order to explore the effectiveness of back-porpagation( BP) neural network shear capacity prediction of simply supported reactive powder concrete( RPC) beams,the shearing failure test results of the 14high-strength re...
Indoor Fingerprint Localization Method Using Channel State Information
Received signal strength indicator ( RSSI) is always used to generate fingerprint in conventional indoor fingerprint positioning technology. However, RSSI is vulnerable to environmental changing, and the positioning accu...
Experiment on Reducing Noise of Reducer in Electric Forklift
Aiming at the loud howlround of the reducer in a certain type of electric forklift,which affected the qualified rate of assembly,the formation mechanism of howlround of the reducer was analyzed. The pinion of the reducer...
Photocatalyzed Degradation of Acid Yellow G via Nano-ZnO Loaded Bentonite
1
Combined 16S rDNA and gyrB Gene Sequence to Identify a Strain of Bacillus cereus
The strain YSSNJ-005 isolated from the bacillus coagulans was used as objective strain.16 S rD NA and gyrB gene sequence were combined to identify the strain exactly.The results show that the strain YSSNJ-005 is prelimin...