void HMM::setEmissionPropability(array1D mean, array1D std){
assert(mean.size() >= this->stateCount);
assert(std.size() >= this->stateCount);
for (unsigned int i = 0; i < this->stateCount; i += 1){
this->setEmissionPropability(mean[i], std[i], i);
}
}
void simulateHMM(
array1D &data,
std::vector<unsigned int> &state,
const array2D transition,
const std::vector<GaussState> states,
const unsigned int dataSize
){
const unsigned int statesCount = states.size();
data.resize(dataSize);
state.resize(dataSize);
state[0] = std::rand() % statesCount;
data[0] = states[state[0]].random();
for (unsigned int t = 1; t < dataSize; t += 1){
double r = (double) rand() / (double) RAND_MAX;
unsigned int s;
for (s = 0; s < statesCount; s += 1){
r -= transition[state[t - 1]][s];
if (r <= 0){
break;
}
}
if (s == 3){
s = 2;
}
state[t] = s;
data[t] = states[state[t]].random();
}
};
double hiddenMarkovModel::normaliseArray(array1D &arr){
double sum = 0;
for (unsigned int i = 0, end = arr.size(); i < end; i += 1){
sum += arr[i];
}
rescaleArray(arr, sum);
return sum;
};
void hiddenMarkovModel::rescaleArray(array1D &arr, double factor){
for (unsigned int i = 0, end = arr.size(); i < end; i += 1){
arr[i] /= factor;
}
};
void HMM::getBinningRange(array1D &range){
range.resize(2);
range[0] = this->binner->getMin();
range[1] = this->binner->getMax();
}
