forked from kscience/kmath
89 lines
3.3 KiB
Kotlin
89 lines
3.3 KiB
Kotlin
|
/*
|
||
|
|
* Copyright 2015 Alexander Nozik.
|
||
|
|
*
|
||
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
||
|
|
* you may not use this file except in compliance with the License.
|
||
|
|
* You may obtain a copy of the License at
|
||
|
|
*
|
||
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
||
|
|
*
|
||
|
|
* Unless required by applicable law or agreed to in writing, software
|
||
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
||
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
||
|
|
* See the License for the specific language governing permissions and
|
||
|
|
* limitations under the License.
|
||
|
|
*/
|
||
|
|
package ru.inr.mass.minuit
|
||
|
|
|
||
|
|
import space.kscience.kmath.optimization.minuit.MINUITPlugin
|
||
|
|
|
||
|
|
/**
|
||
|
|
*
|
||
|
|
* @version $Id$
|
||
|
|
*/
|
||
|
|
internal object MnPosDef {
|
||
|
|
fun test(st: MinimumState, prec: MnMachinePrecision): MinimumState {
|
||
|
|
val err: MinimumError = test(st.error(), prec)
|
||
|
|
return MinimumState(st.parameters(), err, st.gradient(), st.edm(), st.nfcn())
|
||
|
|
}
|
||
|
|
|
||
|
|
fun test(e: MinimumError, prec: MnMachinePrecision): MinimumError {
|
||
|
|
val err: MnAlgebraicSymMatrix = e.invHessian().copy()
|
||
|
|
if (err.size() === 1 && err[0, 0] < prec.eps()) {
|
||
|
|
err[0, 0] = 1.0
|
||
|
|
return MinimumError(err, MnMadePosDef())
|
||
|
|
}
|
||
|
|
if (err.size() === 1 && err[0, 0] > prec.eps()) {
|
||
|
|
return e
|
||
|
|
}
|
||
|
|
// std::cout<<"MnPosDef init matrix= "<<err<<std::endl;
|
||
|
|
val epspdf: Double = max(1e-6, prec.eps2())
|
||
|
|
var dgmin: Double = err[0, 0]
|
||
|
|
for (i in 0 until err.nrow()) {
|
||
|
|
if (err[i, i] < prec.eps2()) {
|
||
|
|
MINUITPlugin.logStatic("negative or zero diagonal element $i in covariance matrix")
|
||
|
|
}
|
||
|
|
if (err[i, i] < dgmin) {
|
||
|
|
dgmin = err[i, i]
|
||
|
|
}
|
||
|
|
}
|
||
|
|
var dg = 0.0
|
||
|
|
if (dgmin < prec.eps2()) {
|
||
|
|
dg = 1.0 + epspdf - dgmin
|
||
|
|
// dg = 0.5*(1. + epspdf - dgmin);
|
||
|
|
MINUITPlugin.logStatic("added $dg to diagonal of error matrix")
|
||
|
|
}
|
||
|
|
val s: RealVector = ArrayRealVector(err.nrow())
|
||
|
|
val p = MnAlgebraicSymMatrix(err.nrow())
|
||
|
|
for (i in 0 until err.nrow()) {
|
||
|
|
err[i, i] = err[i, i] + dg
|
||
|
|
if (err[i, i] < 0.0) {
|
||
|
|
err[i, i] = 1.0
|
||
|
|
}
|
||
|
|
s.setEntry(i, 1.0 / sqrt(err[i, i]))
|
||
|
|
for (j in 0..i) {
|
||
|
|
p[i, j] = err[i, j] * s.getEntry(i) * s.getEntry(j)
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
// std::cout<<"MnPosDef p: "<<p<<std::endl;
|
||
|
|
val eval: RealVector = p.eigenvalues()
|
||
|
|
val pmin: Double = eval.getEntry(0)
|
||
|
|
var pmax: Double = eval.getEntry(eval.getDimension() - 1)
|
||
|
|
// std::cout<<"pmin= "<<pmin<<" pmax= "<<pmax<<std::endl;
|
||
|
|
pmax = max(abs(pmax), 1.0)
|
||
|
|
if (pmin > epspdf * pmax) {
|
||
|
|
return e
|
||
|
|
}
|
||
|
|
val padd = 0.001 * pmax - pmin
|
||
|
|
MINUITPlugin.logStatic("eigenvalues: ")
|
||
|
|
for (i in 0 until err.nrow()) {
|
||
|
|
err[i, i] = err[i, i] * (1.0 + padd)
|
||
|
|
MINUITPlugin.logStatic(java.lang.Double.toString(eval.getEntry(i)))
|
||
|
|
}
|
||
|
|
// std::cout<<"MnPosDef final matrix: "<<err<<std::endl;
|
||
|
|
MINUITPlugin.logStatic("matrix forced pos-def by adding $padd to diagonal")
|
||
|
|
// std::cout<<"eigenvalues: "<<eval<<std::endl;
|
||
|
|
return MinimumError(err, MnMadePosDef())
|
||
|
|
}
|
||
|
|
}
|