public static MultisetKmer <Ty> ToMultisetVarKmer <Ty>(this IEnumerable <DiscreteEventSeries <Ty> > series, int k)
{
MultisetKmer <Ty> multiset = new MultisetKmer <Ty>(k);
series.ForEach(item => multiset.AddDiscreteEventSeriesVarKmer(item, k));
return(multiset);
}
public static MultisetKmer <Ty> ToMultisetVarKmer <Ty>(this DiscreteEventSeries <Ty> series, int k)
{
MultisetKmer <Ty> multiset = new MultisetKmer <Ty>(k);
multiset.AddDiscreteEventSeriesVarKmer(series, k);
return(multiset);
}
public static MultisetKmer <A> MultisetKmerUnion <A>(this IEnumerable <MultisetKmer <A> > sets)
{
//TODO check they all have the same k?
MultisetKmer <A> d = new MultisetKmer <A>(sets.First().maxK); //sets.Select (multiset => multiset.Count).Max());
//TODO add number
sets.ForEach(aset => aset.ForEach(kvp => d.AddKmer(kvp.Key, kvp.Value)));
return(d);
}
/*
* private IEnumerable<TupleStruct<Kmer<Ty>, double>> ExtractUncharacteristicKmersForClass (int classIndex, MultisetKmer<Ty> thisClass, MultisetKmer<Ty> baseline)
* {
* //This will only work with an enormous amount of data for low k.
* }
*/
//Calculation:
//Synthesize features for an item.
//TODO: Enforce contract
public double[] SynthesizeFeatures(DiscreteEventSeries <Ty> item)
{
double[] vals = new double[kmerCount];
MultisetKmer <Ty> ms = item.ToMultisetVarKmer <Ty>(k);
foreach (KeyValuePair <Kmer <Ty>, int> kvp in ms)
{
int index = 0;
if (kmersOntoIndex.TryGetValue(kvp.Key, out index))
{
vals[index] = kvp.Value / (double)ms.Size((int)kvp.Key.Count);
}
}
return(vals);
}
protected override IEventSeriesScalarRegressor <Ty>[] CreateRegressors(DiscreteSeriesDatabase <Ty> data)
{
//Partition into class and classless groups.
Tuple <IEnumerable <DiscreteEventSeries <Ty> >, IEnumerable <DiscreteEventSeries <Ty> > > partitioned = data.data.Partition(item => item.labels.ContainsKey(ClassificationCriterion));
IEnumerable <DiscreteEventSeries <Ty> > noClass = partitioned.Item2; //This item does not have a class over the category label for which the feature synthezer is being created.
IEnumerable <DiscreteEventSeries <Ty> > inClass = partitioned.Item1;
IEnumerable <IGrouping <string, DiscreteEventSeries <Ty> > > groupings = inClass.GroupBy(item => item.labels[ClassificationCriterion]);
//Establish multisets for each class (parallelized).
Tuple <string, MultisetKmer <Ty> >[] classSets = groupings.AsParallel().Select(grp => Tuple.Create(grp.Key, grp.ToMultisetVarKmer(k))).ToArray(); //Used twice. Make it an array.
//Establish the baseline (all data)
MultisetKmer <Ty> baseline = noClass.ToMultisetVarKmer(k).Cons(classSets.Select(a => a.Item2)).MultisetKmerUnion();
//Create regressors (in parallel).
return(classSets.AsParallel().Select(ntp => new ItemVarKmerFrequencyRegressor <Ty>(ntp.Item1, minSignificantCount, smoothingAmount, featuresToUse, k, baseline, ntp.Item2)).ToArray());
}
/*
* private IEnumerable<TupleStruct<Kmer<Ty>, double>> ExtractUncharacteristicKmersForClass (int classIndex, MultisetKmer<Ty> thisClass, MultisetKmer<Ty> baseline)
* {
* //This will only work with an enormous amount of data for low k.
* }
*/
//Calculation:
//Synthesize features for an item.
//TODO: Enforce contract
public double[] SynthesizeFeatures(DiscreteEventSeries <Ty> item)
{
double[] vals = new double[classCount];
MultisetKmer <Ty> ms = item.ToMultisetVarKmer <Ty>(k);
foreach (KeyValuePair <Kmer <Ty>, int> kvp in ms)
{
Dictionary <int, double> classesWithKvp;
if (learnedCharacteristicKmers.TryGetValue(kvp.Key, out classesWithKvp))
{
//Console.WriteLine ("\tFound kmer " + kvp.Key + ".");
foreach (KeyValuePair <int, double> @class in classesWithKvp)
{
//Console.WriteLine ("\t\tClass " + @class.Key + ", Value " + @class.Value + ", Times " + kvp.Value);
vals[@class.Key] += kvp.Value * @class.Value;
}
}
}
return(vals);
}
public void TrainModelRatios(MultisetKmer <A> baselineClass, MultisetKmer <A> thisClass)
{
List <KeyValuePair <Kmer <A>, double> > rawModel = new List <KeyValuePair <Kmer <A>, double> >();
int totalCount = 0;
foreach (Kmer <A> key in thisClass.Keys)
{
int thisCount = thisClass.getCount(key);
totalCount += thisCount;
if (thisCount > regressor.minSignificantCount)
{
double thisFrac = thisClass.GetKeyFrac(key);
double baseFrac = baselineClass.GetKeyFracLaplace(key, regressor.smoothingAmount);
if (thisFrac > baseFrac)
{
rawModel.Add(key, thisFrac / baseFrac);
}
}
}
regressor.finalizeModel(rawModel, totalCount);
}
//
//Kmer variadic k multiset
//
public static void AddDiscreteEventSeriesVarKmer <Ty>(this MultisetKmer <Ty> multiset, DiscreteEventSeries <Ty> series, int k)
{
Ty[] arr = series.data;
multiset.ConsumeEventSeriesKmer(arr);
}
//Extract characteristic kmers (top n more common than baseline that occur at least q times).
private IEnumerable <TupleStruct <Kmer <Ty>, double> > ExtractCharacteristicKmersForClass(int classIndex, MultisetKmer <Ty> thisClass, MultisetKmer <Ty> baseline)
{
List <TupleStruct <Kmer <Ty>, double> > thisClassCharacteristicKmersStore = new List <TupleStruct <Kmer <Ty>, double> >();
foreach (KeyValuePair <Kmer <Ty>, int> kvp in thisClass)
{
if (kvp.Value > minKmerCount)
{
double thisFreq = kvp.Value / (double)thisClass.Size(kvp.Key.Count);
double baseFreq = baseline.GetKeyFracLaplace(kvp.Key, smoothingAmt);
//Console.WriteLine ("Class: " + classIndex + " Kmer: " + kvp.Value + ", class freq " + thisFreq + ", base freq " + baseFreq);
//TODO: Advanced logic.
if (thisFreq > baseFreq)
{
double kmerValue = thisFreq / baseFreq - 1;
//Console.WriteLine ("Adding kmer " + kvp.Key + " weight " + kmerValue + " for class " + classIndex);
thisClassCharacteristicKmersStore.Add(new TupleStruct <Kmer <Ty>, double>(kvp.Key, kmerValue));
}
}
}
return(thisClassCharacteristicKmersStore.OrderByDescending(tup => Math.Abs(tup.Item2)).Take((int)kmersToTake));
}
public static ClassCharacteristicSetKmer<A> BuildSubtractiveDifference(string name, MultisetKmer<A> baselineClass, MultisetKmer<A> thisClass, uint countCutoff)
{
ClassCharacteristicSetKmer<A> newSet = new ClassCharacteristicSetKmer<A>(name, Math.Min (baselineClass.maxK, thisClass.maxK));
//TODO statistically significant?
//TODO diffence amount?
foreach(Kmer<A> key in thisClass.Keys){
if(thisClass.getCount(key) > countCutoff){
double thisFrac = thisClass.GetKeyFrac(key);
double baseFrac = baselineClass.GetKeyFrac (key);
if(thisFrac > baseFrac){
newSet.Add (key, thisFrac - baseFrac);
}
}
}
//TODO select top x?
//double[] function?
return newSet;
}
public ItemVarKmerFrequencyRegressor(string name, int minSignificantCount, int smoothingAmount, int featuresToUse, int k, MultisetKmer <A> baselineClass, MultisetKmer <A> thisClass) : this(name, minSignificantCount, smoothingAmount, featuresToUse, k)
{
TrainModelRatios(baselineClass, thisClass);
}
