@operator(
value = {"percent_absolute_deviation"},
content_type = IType.FLOAT,
category = {IOperatorCategory.MAP_COMPARAISON},
concept = {IConcept.STATISTIC})
@doc(
value =
"percent absolute deviation indicator for 2 series of values: percent_absolute_deviation(list_vals_observe,list_vals_sim)",
examples = {
@example(
value = "percent_absolute_deviation([200,300,150,150,200],[250,250,100,200,200])",
isExecutable = false)
})
public static double percentAbsoluteDeviation(
final IScope scope, final IList<Double> vals1, final IList<Double> vals2) {
if (vals1 == null || vals2 == null) {
return 1;
}
int nb = vals1.size();
if (nb != vals2.size()) {
return 0;
}
double sum = 0;
double coeff = 0;
for (int i = 0; i < nb; i++) {
double val1 = Cast.asFloat(scope, vals1.get(i));
double val2 = Cast.asFloat(scope, vals2.get(i));
coeff += val1;
sum += FastMath.abs(val1 - val2) * 100.0;
}
if (coeff == 0) {
return 0;
}
return sum / coeff;
}
private static Instances convertToInstances(
final IScope scope,
final IList<String> attributes,
final IAddressableContainer<Integer, IAgent, Integer, IAgent> agents)
throws GamaRuntimeException {
FastVector attribs = new FastVector();
for (String att : attributes) {
attribs.addElement(new Attribute(att));
}
Instances dataset =
new Instances(scope.getAgentScope().getName(), attribs, agents.length(scope));
for (IAgent ag : agents.iterable(scope)) {
int nb = attributes.size();
double vals[] = new double[nb];
for (int i = 0; i < nb; i++) {
String attrib = attributes.get(i);
Double var = Cast.asFloat(scope, ag.getDirectVarValue(scope, attrib));
vals[i] = var;
}
Instance instance = new Instance(1, vals);
dataset.add(instance);
}
return dataset;
}
public void updateXValues(IScope scope, int chartCycle, int targetNb) {
Object xval, xlab;
if (this.useXSource || this.useXLabels) {
if (this.useXSource) {
xval = xsource.resolveAgainst(scope).value(scope);
} else {
xval = xlabels.resolveAgainst(scope).value(scope);
}
if (this.useXLabels) {
xlab = xlabels.resolveAgainst(scope).value(scope);
} else {
xlab = xsource.resolveAgainst(scope).value(scope);
}
if (xval instanceof GamaList) {
IList xv2 = Cast.asList(scope, xval);
IList xl2 = Cast.asList(scope, xlab);
if (this.useXSource && xv2.size() > 0 && xv2.get(0) instanceof Number) {
XSeriesValues = new ArrayList<Double>();
Xcategories = new ArrayList<String>();
for (int i = 0; i < xv2.size(); i++) {
XSeriesValues.add(new Double(Cast.asFloat(scope, xv2.get(i))));
Xcategories.add(Cast.asString(scope, xl2.get(i)));
}
} else {
if (xv2.size() > Xcategories.size()) {
Xcategories = new ArrayList<String>();
for (int i = 0; i < xv2.size(); i++) {
if (i >= XSeriesValues.size()) {
XSeriesValues.add(new Double(getXCycleOrPlusOneForBatch(scope, chartCycle)));
}
Xcategories.add(Cast.asString(scope, xl2.get(i)));
}
}
}
if (xv2.size() < targetNb) {
throw GamaRuntimeException.error(
"The x-serie length ("
+ xv2.size()
+ ") should NOT be shorter than any series length ("
+ targetNb
+ ") !",
scope);
}
} else {
if (this.useXSource && xval instanceof Number) {
double dvalue = Cast.asFloat(scope, xval);
String lvalue = Cast.asString(scope, xlab);
XSeriesValues.add(new Double(dvalue));
Xcategories.add(lvalue);
}
if (targetNb == -1 && !this.forceNoXAccumulate) targetNb = XSeriesValues.size() + 1;
while (XSeriesValues.size() < targetNb) {
XSeriesValues.add(new Double(getXCycleOrPlusOneForBatch(scope, chartCycle)));
Xcategories.add(Cast.asString(scope, xlab));
}
}
}
if (!this.useXSource && !this.useXLabels) {
if (targetNb == -1 && !this.forceNoXAccumulate && commonXindex >= XSeriesValues.size())
targetNb = XSeriesValues.size() + 1;
while (XSeriesValues.size() < targetNb) {
addCommonXValue(scope, getXCycleOrPlusOneForBatch(scope, chartCycle));
}
}
}
@operator(
value = {"kappa"},
content_type = IType.FLOAT,
category = {IOperatorCategory.MAP_COMPARAISON},
concept = {})
@doc(
value =
"kappa indicator for 2 map comparisons: kappa(list_vals1,list_vals2,categories, weights). Reference: Cohen, J. A coefficient of agreement for nominal scales. Educ. Psychol. Meas. 1960, 20. ",
examples = {
@example(
value =
"kappa([cat1,cat1,cat2,cat3,cat2],[cat2,cat1,cat2,cat1,cat2],[cat1,cat2,cat3], [1.0, 2.0, 3.0, 1.0, 5.0])",
isExecutable = false)
})
public static double kappa(
final IScope scope,
final IList<Object> vals1,
final IList<Object> vals2,
final List<Object> categories,
final IList<Object> weights) {
if (vals1 == null || vals2 == null) {
return 1;
}
int nb = vals1.size();
if (nb != vals2.size()) {
return 0;
}
int nbCat = categories.size();
double[] X = new double[nbCat];
double[] Y = new double[nbCat];
double[][] contigency = new double[nbCat][nbCat];
for (int j = 0; j < nbCat; j++) {
X[j] = 0;
Y[j] = 0;
for (int k = 0; k < nbCat; k++) {
contigency[j][k] = 0;
}
}
Map<Object, Integer> categoriesId = new TOrderedHashMap<Object, Integer>();
for (int i = 0; i < nbCat; i++) {
categoriesId.put(categories.get(i), i);
}
double total = 0;
for (int i = 0; i < nb; i++) {
double weight = weights == null ? 1.0 : Cast.asFloat(scope, weights.get(i));
total += weight;
Object val1 = vals1.get(i);
Object val2 = vals2.get(i);
int indexVal1 = categoriesId.get(val1);
int indexVal2 = categoriesId.get(val2);
X[indexVal1] += weight;
Y[indexVal2] += weight;
contigency[indexVal1][indexVal2] += weight;
}
for (int j = 0; j < nbCat; j++) {
X[j] /= total;
Y[j] /= total;
for (int k = 0; k < nbCat; k++) {
contigency[j][k] /= total;
}
}
double po = 0;
double pe = 0;
for (int i = 0; i < nbCat; i++) {
po += contigency[i][i];
pe += X[i] * Y[i];
}
if (pe == 1) {
return 1;
}
return (po - pe) / (1 - pe);
}
@operator(
value = {"kappa_sim"},
content_type = IType.FLOAT,
category = {IOperatorCategory.MAP_COMPARAISON},
concept = {})
@doc(
value =
"kappa simulation indicator for 2 map comparisons: kappa(list_valsInits,list_valsObs,list_valsSim, categories, weights). Reference: van Vliet, J., Bregt, A.K. & Hagen-Zanker, A. (2011). Revisiting Kappa to account for change in the accuracy assessment of land-use change models, Ecological Modelling 222(8)",
examples = {
@example(
value =
"kappa([cat1,cat1,cat2,cat2,cat2],[cat2,cat1,cat2,cat1,cat3],[cat2,cat1,cat2,cat3,cat3], [cat1,cat2,cat3],[1.0, 2.0, 3.0, 1.0, 5.0])",
isExecutable = false)
})
public static double kappaSimulation(
final IScope scope,
final IList<Object> valsInit,
final IList<Object> valsObs,
final IList<Object> valsSim,
final List<Object> categories,
final IList<Object> weights) {
if (valsInit == null || valsObs == null || valsSim == null) {
return 1;
}
int nb = valsInit.size();
if (nb != valsObs.size() || nb != valsSim.size()) {
return 0;
}
int nbCat = categories.size();
double[] O = new double[nbCat];
double[][] contigency = new double[nbCat][nbCat];
double[][] contigencyOA = new double[nbCat][nbCat];
double[][] contigencyOS = new double[nbCat][nbCat];
for (int j = 0; j < nbCat; j++) {
O[j] = 0;
for (int k = 0; k < nbCat; k++) {
contigency[j][k] = 0;
contigencyOA[j][k] = 0;
contigencyOS[j][k] = 0;
}
}
Map<Object, Integer> categoriesId = new TOrderedHashMap<Object, Integer>();
for (int i = 0; i < nbCat; i++) {
categoriesId.put(categories.get(i), i);
}
double total = 0;
for (int i = 0; i < nb; i++) {
double weight = weights == null ? 1.0 : Cast.asFloat(scope, weights.get(i));
total += weight;
Object val1 = valsObs.get(i);
Object val2 = valsSim.get(i);
Object valO = valsInit.get(i);
int indexVal1 = categoriesId.get(val1);
int indexVal2 = categoriesId.get(val2);
int indexValO = categoriesId.get(valO);
O[indexValO] += weight;
contigency[indexVal1][indexVal2] += weight;
contigencyOA[indexValO][indexVal1] += weight;
contigencyOS[indexValO][indexVal2] += weight;
}
for (int j = 0; j < nbCat; j++) {
for (int k = 0; k < nbCat; k++) {
contigency[j][k] /= total;
if (O[j] > 0) {
contigencyOA[j][k] /= O[j];
contigencyOS[j][k] /= O[j];
}
}
O[j] /= total;
}
double po = 0;
double pe = 0;
for (int j = 0; j < nbCat; j++) {
po += contigency[j][j];
double sum = 0;
for (int i = 0; i < nbCat; i++) {
sum += contigencyOA[j][i] * contigencyOS[j][i];
}
pe += O[j] * sum;
}
if (pe == 1) {
return 1;
}
return (po - pe) / (1 - pe);
}