vector computation
1.Girosi, F. "An Equivalence between Sparse Approximation and Support Vector Machines." Neural Computation 10 (1998): 1455-1480.
2.This paper presents the overall structure, design idea and systematic function and application condition of information analysis and management system on hydrology and sediment of the Yangtze river and mainly introduces function realization of 8 sub-systems of editing sub-system of vector data, management discipline of object relation database, computation sub-system of hydrology and sediment sub-system, visualized analysis sub-system of water and sediment information, analysis sub-system of the Yangtze river channel evolution, inquiry sub-system of fundamental information, 3 dimensional visualized sub-system of the Yangtze river, network publication sub-system of hydrology, sediment and river channel information.
3.When a vector assignment statement such as A= B + C is executed on a SIMD computer, the actual result of computation sometimes does not correspond to the ordinary concept of vector operations.
4.Matrix and vector template classes are defined based upon object-oriented technology(OOT).The conventional computation among matrix,vector and number is realized by overriding operators.According to the property of matrix,the scanty reserve way of object data is used,and the computation between matrixes with different reserve way is realized.The template classes is versatile,the program codes can be effectively reused,maintained and extended.
5.Meanwhile, two numerical computation object(including class Matrix Vector) are defined for the convenience of the programming.
6.Support Vector Machine (SVM) was invented by Boser, Guyon and Vapnik, and was put forward at the 5th annual ACM Workshop (1992) on Computation Learning Theory (COLT) for the first time.
7.This paper introduces loop optimization in a compiler for the vector extension of a language based on FORTRAN, ANSI X3.9—1978 (ISO 1539—1980). Also studied is the approach to the computation of the head, increment, and length of vectors.
8.The result showed that, with the equivalent substitution method of kinematic joint for error analysis, its computer-aided analysis was realized and the malady of tedium and complexity with the analysis and computation of the vector loop was eliminated, making the error analysis of the mechanism simple, easy, and reliable.
