public bool EndsWith(Ngram <TItem> items, Direction dir, IEqualityComparer <TItem> comparer)
{
if (items.Length > Length)
{
return(false);
}
IEnumerable <TItem> x = items;
IEnumerable <TItem> y = _items;
if (dir == Direction.LeftToRight)
{
x = x.Reverse();
y = y.Reverse();
}
foreach (Tuple <TItem, TItem> item in x.Zip(y))
{
if (!comparer.Equals(item.Item1, item.Item2))
{
return(false);
}
}
return(true);
}
public double GetProbability(TItem item, Ngram <TItem> context)
{
if (context.Length != _ngramSize - 1)
{
throw new ArgumentException("The context size is not valid.", "context");
}
return(_smoother.GetProbability(item, context));
}
public double GetProbability(TItem item, Ngram <TItem> context)
{
FrequencyDistribution <TItem> fd = _cfd[context];
if (fd.SampleOutcomeCount == 0)
{
return(0);
}
return((double)fd[item] / fd.SampleOutcomeCount);
}
public double GetProbability(TItem item, Ngram <TItem> context)
{
FrequencyDistribution <TItem> freqDist = _cfd[context];
if (freqDist.ObservedSamples.Count == 0)
{
return(0);
}
double numer = freqDist[item] + (freqDist.ObservedSamples.Count * (_lowerOrderModel == null ? 1.0 / freqDist.ObservedSamples.Count
: _lowerOrderModel.GetProbability(item, context.SkipFirst(_dir))));
double denom = freqDist.SampleOutcomeCount + freqDist.ObservedSamples.Count;
return(numer / denom);
}
public NgramModel(int ngramSize, TSeq[] sequences, Func <TSeq, IEnumerable <TItem> > itemsSelector, Direction dir, INgramModelSmoother <TSeq, TItem> smoother)
{
_ngramSize = ngramSize;
_dir = dir;
_smoother = smoother;
_ngrams = new HashSet <Ngram <TItem> >();
var cfd = new ConditionalFrequencyDistribution <Ngram <TItem>, TItem>();
foreach (TSeq seq in sequences)
{
TItem[] items = itemsSelector(seq).ToArray();
for (int i = 0; i <= items.Length - ngramSize; i++)
{
var ngram = new Ngram <TItem>(Enumerable.Range(i, _ngramSize).Select(j => items[j]));
_ngrams.Add(ngram);
Ngram <TItem> context = ngram.TakeAllExceptLast(dir);
TItem item = ngram.GetLast(dir);
cfd[context].Increment(item);
}
}
_smoother.Smooth(ngramSize, sequences, itemsSelector, dir, cfd);
}
public double GetProbability(TItem item, Ngram <TItem> context)
{
FrequencyDistribution <TItem> freqDist = _cfd[context];
if (freqDist.ObservedSamples.Count == 0)
{
return(0);
}
if (context.Length == 0)
{
return((double)freqDist[item] / freqDist.SampleOutcomeCount);
}
int count = freqDist[item];
Tuple <int, int, int> bigN = _bigNs[context];
double gamma = ((_discount1 * bigN.Item1) + (_discount2 * bigN.Item2) + (_discount3 * bigN.Item3)) / freqDist.SampleOutcomeCount;
double d = 0;
if (count == 1)
{
d = _discount1;
}
else if (count == 2)
{
d = _discount2;
}
else if (count > 2)
{
d = _discount3;
}
double prob = (count - d) / freqDist.SampleOutcomeCount;
return(prob + (gamma * _lowerOrderModel.GetProbability(item, context.SkipFirst(_dir))));
}
public bool Equals(Ngram <TItem> other)
{
return(other != null && _hashCode == other._hashCode && _items.SequenceEqual(other._items));
}
public bool EndsWith(Ngram <TItem> items, Direction dir)
{
return(EndsWith(items, dir, EqualityComparer <TItem> .Default));
}
public bool EndsWith(Ngram <TItem> items)
{
return(EndsWith(items, Direction.LeftToRight));
}
public Ngram <TItem> Concat(Ngram <TItem> ngram, Direction dir)
{
return(new Ngram <TItem>(dir == Direction.LeftToRight ? _items.Concat(ngram) : ngram.Concat(_items)));
}
public Ngram <TItem> Concat(Ngram <TItem> ngram)
{
return(Concat(ngram, Direction.LeftToRight));
}
