📄 pacematrix.html
字号:
</DL><HR><A NAME="g1(double, double)"><!-- --></A><H3>g1</H3><PRE>public double[] <B>g1</B>(double a, double b)</PRE><DL><DD>Constructs the Givens rotation<P><DD><DL><DT><B>Parameters:</B><DD><CODE>a</CODE> - <DD><CODE>b</CODE> - <DT><B>Returns:</B><DD>a double array that stores the cosine and sine values</DL></DD></DL><HR><A NAME="g2(double[], int, int, int)"><!-- --></A><H3>g2</H3><PRE>public void <B>g2</B>(double[] cs, int i0, int i1, int j)</PRE><DL><DD><DL></DL></DD></DL><HR><A NAME="forward(weka.classifiers.functions.pace.PaceMatrix, weka.classifiers.functions.pace.IntVector, int)"><!-- --></A><H3>forward</H3><PRE>public void <B>forward</B>(<A HREF="../../../../weka/classifiers/functions/pace/PaceMatrix.html" title="class in weka.classifiers.functions.pace">PaceMatrix</A> b, <A HREF="../../../../weka/classifiers/functions/pace/IntVector.html" title="class in weka.classifiers.functions.pace">IntVector</A> pvt, int k0)</PRE><DL><DD>Forward ordering of columns in terms of response explanation. On input, matrices A and b are already QR-transformed. The indices of transformed columns are stored in the pivoting vector.<P><DD><DL><DT><B>Parameters:</B><DD><CODE>b</CODE> - the PaceMatrix b<DD><CODE>pvt</CODE> - the pivoting vector<DD><CODE>k0</CODE> - the first k0 columns (in pvt) of A are not to be changed</DL></DD></DL><HR><A NAME="mostExplainingColumn(weka.classifiers.functions.pace.PaceMatrix, weka.classifiers.functions.pace.IntVector, int)"><!-- --></A><H3>mostExplainingColumn</H3><PRE>public int <B>mostExplainingColumn</B>(<A HREF="../../../../weka/classifiers/functions/pace/PaceMatrix.html" title="class in weka.classifiers.functions.pace">PaceMatrix</A> b, <A HREF="../../../../weka/classifiers/functions/pace/IntVector.html" title="class in weka.classifiers.functions.pace">IntVector</A> pvt, int ks)</PRE><DL><DD>Returns the index of the column that has the largest (squared) response, when each of columns pvt[ks:] is moved to become the ks-th column. On input, A and b are both QR-transformed.<P><DD><DL><DT><B>Parameters:</B><DD><CODE>b</CODE> - response<DD><CODE>pvt</CODE> - pivoting index of A<DD><CODE>ks</CODE> - columns pvt[ks:] of A are to be tested</DL></DD></DL><HR><A NAME="backward(weka.classifiers.functions.pace.PaceMatrix, weka.classifiers.functions.pace.IntVector, int, int)"><!-- --></A><H3>backward</H3><PRE>public void <B>backward</B>(<A HREF="../../../../weka/classifiers/functions/pace/PaceMatrix.html" title="class in weka.classifiers.functions.pace">PaceMatrix</A> b, <A HREF="../../../../weka/classifiers/functions/pace/IntVector.html" title="class in weka.classifiers.functions.pace">IntVector</A> pvt, int ks, int k0)</PRE><DL><DD>Backward ordering of columns in terms of response explanation. On input, matrices A and b are already QR-transformed. The indices of transformed columns are stored in the pivoting vector. A and b must have the same number of rows, being the (pseudo-)rank.<P><DD><DL><DT><B>Parameters:</B><DD><CODE>b</CODE> - PaceMatrix b<DD><CODE>pvt</CODE> - pivoting vector<DD><CODE>ks</CODE> - number of QR-transformed columns; psuedo-rank of A<DD><CODE>k0</CODE> - first k0 columns in pvt[] are not to be ordered.<DT><B>Returns:</B><DD>void</DL></DD></DL><HR><A NAME="leastExplainingColumn(weka.classifiers.functions.pace.PaceMatrix, weka.classifiers.functions.pace.IntVector, int, int)"><!-- --></A><H3>leastExplainingColumn</H3><PRE>public int <B>leastExplainingColumn</B>(<A HREF="../../../../weka/classifiers/functions/pace/PaceMatrix.html" title="class in weka.classifiers.functions.pace">PaceMatrix</A> b, <A HREF="../../../../weka/classifiers/functions/pace/IntVector.html" title="class in weka.classifiers.functions.pace">IntVector</A> pvt, int ks, int k0)</PRE><DL><DD>Returns the index of the column that has the smallest (squared) response, when the column is moved to become the (ks-1)-th column. On input, A and b are both QR-transformed.<P><DD><DL><DT><B>Parameters:</B><DD><CODE>b</CODE> - response<DD><CODE>pvt</CODE> - pivoting index of A<DD><CODE>ks</CODE> - psudo-rank of A<DD><CODE>k0</CODE> - A[][pvt[0:(k0-1)]] are excluded from the testing.</DL></DD></DL><HR><A NAME="columnResponseExplanation(weka.classifiers.functions.pace.PaceMatrix, weka.classifiers.functions.pace.IntVector, int, int)"><!-- --></A><H3>columnResponseExplanation</H3><PRE>public double <B>columnResponseExplanation</B>(<A HREF="../../../../weka/classifiers/functions/pace/PaceMatrix.html" title="class in weka.classifiers.functions.pace">PaceMatrix</A> b, <A HREF="../../../../weka/classifiers/functions/pace/IntVector.html" title="class in weka.classifiers.functions.pace">IntVector</A> pvt, int j, int ks)</PRE><DL><DD>Returns the squared ks-th response value if the j-th column becomes the ks-th after orthogonal transformation. A[][pvt[ks:j]] (or A[][pvt[j:ks]], if ks > j) and b[] are already QR-transformed on input and will remain unchanged on output. More generally, it returns the inner product of the corresponding row vector of the response PaceMatrix. (To be implemented.)<P><DD><DL><DT><B>Parameters:</B><DD><CODE>b</CODE> - PaceMatrix b<DD><CODE>pvt</CODE> - pivoting vector<DD><CODE>j</CODE> - the column A[pvt[j]][] is to be moved<DD><CODE>ks</CODE> - the target column A[pvt[ks]][]<DT><B>Returns:</B><DD>the squared response value</DL></DD></DL><HR><A NAME="lsqr(weka.classifiers.functions.pace.PaceMatrix, weka.classifiers.functions.pace.IntVector, int)"><!-- --></A><H3>lsqr</H3><PRE>public void <B>lsqr</B>(<A HREF="../../../../weka/classifiers/functions/pace/PaceMatrix.html" title="class in weka.classifiers.functions.pace">PaceMatrix</A> b, <A HREF="../../../../weka/classifiers/functions/pace/IntVector.html" title="class in weka.classifiers.functions.pace">IntVector</A> pvt, int k0)</PRE><DL><DD>QR transformation for a least squares problem A x = b<P><DD><DL><DT><B>Parameters:</B><DD><CODE>b</CODE> - PaceMatrix b<DD><CODE>pvt</CODE> - pivoting vector<DD><CODE>k0</CODE> - the first k0 columns of A (indexed by pvt) are pre-chosen. (But subject to rank examination.) For example, the constant term may be reserved, in which case k0 = 1.<DT><B>Returns:</B><DD>void; implicitly both A and b are transformed. pvt.size() is is the psuedo-rank of A.</DL></DD></DL><HR><A NAME="lsqrSelection(weka.classifiers.functions.pace.PaceMatrix, weka.classifiers.functions.pace.IntVector, int)"><!-- --></A><H3>lsqrSelection</H3><PRE>public void <B>lsqrSelection</B>(<A HREF="../../../../weka/classifiers/functions/pace/PaceMatrix.html" title="class in weka.classifiers.functions.pace">PaceMatrix</A> b, <A HREF="../../../../weka/classifiers/functions/pace/IntVector.html" title="class in weka.classifiers.functions.pace">IntVector</A> pvt, int k0)</PRE><DL><DD>QR transformation for a least squares problem A x = b<P><DD><DL><DT><B>Parameters:</B><DD><CODE>b</CODE> - PaceMatrix b<DD><CODE>pvt</CODE> - pivoting vector<DD><CODE>k0</CODE> - the first k0 columns of A (indexed by pvt) are pre-chosen. (But subject to rank examination.) For example, the constant term may be reserved, in which case k0 = 1.<DT><B>Returns:</B><DD>void; implicitly both A and b are transformed. pvt.size() is is the psuedo-rank of A.</DL></DD></DL><HR><A NAME="positiveDiagonal(weka.classifiers.functions.pace.PaceMatrix, weka.classifiers.functions.pace.IntVector)"><!-- --></A><H3>positiveDiagonal</H3><PRE>public void <B>positiveDiagonal</B>(<A HREF="../../../../weka/classifiers/functions/pace/PaceMatrix.html" title="class in weka.classifiers.functions.pace">PaceMatrix</A> Y, <A HREF="../../../../weka/classifiers/functions/pace/IntVector.html" title="class in weka.classifiers.functions.pace">IntVector</A> pvt)</PRE><DL><DD>Sets all diagonal elements to be positive (or nonnegative) without changing the least squares solution<P><DD><DL><DT><B>Parameters:</B><DD><CODE>Y</CODE> - the response<DD><CODE>pvt</CODE> - the pivoted column index</DL></DD></DL><HR><A NAME="steplsqr(weka.classifiers.functions.pace.PaceMatrix, weka.classifiers.functions.pace.IntVector, int, int, boolean)"><!-- --></A><H3>steplsqr</H3><PRE>public void <B>steplsqr</B>(<A HREF="../../../../weka/classifiers/functions/pace/PaceMatrix.html" title="class in weka.classifiers.functions.pace">PaceMatrix</A> b, <A HREF="../../../../weka/classifiers/functions/pace/IntVector.html" title="class in weka.classifiers.functions.pace">IntVector</A> pvt, int ks, int j, boolean adjoin)</PRE><DL><DD>Stepwise least squares QR-decomposition of the problem A x = b<P><DD><DL><DT><B>Parameters:</B><DD><CODE>b</CODE> - PaceMatrix b<DD><CODE>pvt</CODE> - pivoting vector<DD><CODE>ks</CODE> - number of transformed columns<DD><CODE>j</CODE> - pvt[j], the column to adjoin or delete<DD><CODE>adjoin</CODE> - to adjoin if true; otherwise, to delete</DL></DD></DL><HR><A NAME="rsolve(weka.classifiers.functions.pace.PaceMatrix, weka.classifiers.functions.pace.IntVector, int)"><!-- --></A><H3>rsolve</H3><PRE>public void <B>rsolve</B>(<A HREF="../../../../weka/classifiers/functions/pace/PaceMatrix.html" title="class in weka.classifiers.functions.pace">PaceMatrix</A> b, <A HREF="../../../../weka/classifiers/functions/pace/IntVector.html" title="class in weka.classifiers.functions.pace">IntVector</A> pvt, int kp)</PRE><DL><DD>Solves upper-triangular equation R x = b<P><DD><DL><DT><B>Parameters:</B><DD><CODE>b</CODE> - the response<DD><CODE>pvt</CODE> - the pivoting vector<DD><CODE>kp</CODE> - the number of the first columns involved<DT><B>Returns:</B><DD>void. On output, the solution is stored in b</DL></DD></DL><HR><A NAME="rbind(weka.classifiers.functions.pace.PaceMatrix)"><!-- --></A><H3>rbind</H3><PRE>public <A HREF="../../../../weka/classifiers/functions/pace/PaceMatrix.html" title="class in weka.classifiers.functions.pace">PaceMatrix</A> <B>rbind</B>(<A HREF="../../../../weka/classifiers/functions/pace/PaceMatrix.html" title="class in weka.classifiers.functions.pace">PaceMatrix</A> b)</PRE><DL><DD>Returns a new matrix which binds two matrices together with rows.<P><DD><DL><DT><B>Parameters:</B><DD><CODE>b</CODE> - the second matrix<DT><B>Returns:</B><DD>the combined matrix</DL></DD></DL><HR><A NAME="cbind(weka.classifiers.functions.pace.PaceMatrix)"><!-- --></A><H3>cbind</H3><PRE>public <A HREF="../../../../weka/classifiers/functions/pace/PaceMatrix.html" title="class in weka.classifiers.functions.pace">PaceMatrix</A> <B>cbind</B>(<A HREF="../../../../weka/classifiers/functions/pace/PaceMatrix.html" title="class in weka.classifiers.functions.pace">PaceMatrix</A> b)</PRE><DL><DD>Returns a new matrix which binds two matrices with columns.<P><DD><DL><DT><B>Parameters:</B><DD><CODE>b</CODE> - the second matrix<DT><B>Returns:</B><DD>the combined matrix</DL></DD></DL><HR><A NAME="nnls(weka.classifiers.functions.pace.PaceMatrix, weka.classifiers.functions.pace.IntVector)"><!-- --></A><H3>nnls</H3><PRE>public <A HREF="../../../../weka/classifiers/functions/pace/DoubleVector.html" title="class in weka.classifiers.functions.pace">DoubleVector</A> <B>nnls</B>(<A HREF="../../../../weka/classifiers/functions/pace/PaceMatrix.html" title="class in weka.classifiers.functions.pace">PaceMatrix</A> b, <A HREF="../../../../weka/classifiers/functions/pace/IntVector.html" title="class in weka.classifiers.functions.pace">IntVector</A> pvt)</PRE><DL><DD>Solves the nonnegative linear squares problem. That is, <p> <center> min || A x - b||
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -