Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Solve Integral Equation by Basis Function Expansions
The eigenfunction is an integral function and difficult to solve in closed form. A general
strategy for solving the eigenfunction problem in (4) is to convert the continuous eigen-analysis
problem to an appropriate discrete eigen-analysis task [24]. In this paper, we use basis function
expansion methods to achieve this conversion.
Let { j (t )} be the series of Fourier functions. For each j, define 2j-1 2 j 2j. We
expand each genetic variant profile X i (t ) as a linear combination of the basis function j :
T
X i (t ) Cij j (t ).
(7)
j 1
Define the vector-valued function X (t ) [ X 1 (t ), , X N (t )]T and the vector-valued
function (t ) [1 (t ),,T (t )]T . The joint expansion of all N genetic variant profiles can be
expressed as
X (t ) C (t ) .
(8)
where the matrix C is given by
C11, C1T
C
CN 1 CNT .
.
In matrix form we can express the variance-covariance function of the genetic variant profiles as
1 T
X ( s ) X (t )
N
1
T ( s )C T C (t ).
N
R ( s, t )
(9)
Similarly, the eigenfunction (t ) can be expanded as
T
T
j 1
j1
(t ) b j j (t ) and D4(t ) 4j b j j (t ) or
(t ) (t )T b and D 4(t ) (t )T S 0 b
(10)
1
1
where b [b1 ,..., bT ]T and S0 diag ( 14 ,..., T4 ) . Let S diag ((1 14 ) 2 ,..., (1 T4 ) 2 ). Then,
we have
(t ) D 4(t ) (t )T S 2 b.
(11)
Substituting expansions (9) and (11) of variance-covariance R(s,t) and eigenfunction (t ) into
the functional eigenequation (6), we obtain
(t ) T
1 T
C Cb T (t ) S 2 b ,
N
(12)
Since equation (12) must hold for all t, we obtain the following eigenequation:
1 T
C Cb S 2 b ,
N
(13)
which can be rewritten as
[S (
S(
1 T
C C ) S ][ S 1b] [ S 1b] , or
N
1 T
C C ) Su u ,
N
(14)
where u S 1b . Thus, b Su and (t ) (t )T b is a solution to eigenequation (6).
Note that u j , u j 1 and u j , uk 0, for k j. Therefore, we obtain a set of orthonormal
eigenfunctions with an inner product of two functions defined in equation (4), as shown in
equation (15):
|| j ||2 bTj S 2b j uTj SS 2 Su j 1 and j ,k bTj S 2bk uTj uk 0 .
(15)