Class Model

pattern/vector/__init__.py:768–1200

KMEANS, HIERARCHICAL, ALL = "k-means", "hierarchical", "all"

class Model(object):
    
    def __init__(self, documents=[], weight=TFIDF):
        """ A model is a bag-of-word representation of a corpus of documents, 
            where each document vector is a bag of (word, relevance)-items.
            Vectors can then be compared for similarity using a distance metric.
            The weighting scheme can be: relative TF, TFIDF (default), BINARY, None,
            where None means that the original weights are used.
        """
        self.description = ""             # Description of the dataset: author e-mail, etc.
        self._documents  = readonlylist() # List of documents (read-only).
        self._index      = {}             # Document.name => Document
        self._df         = {}             # Cache of document frequency per word.
        self._cos        = {}             # Cache of ((d1.id, d2.id), relevance)-items (cosine similarity).
        self._ig         = {}             # Cache of (word, information gain)-items.
        self._vector     = None           # Cache of model vector with all the features in the model.
        self._lsa        = None           # LSA matrix with reduced dimensionality.
        self._weight     = weight         # Weight used in Document.vector (TF, TFIDF, BINARY or None).
        self._update()
        self.extend(documents)
    
    @property
    def documents(self):
        return self._documents

    @property
    def terms(self):
        return self.vector.keys()
        
    features = words = terms
    
    @property
    def classes(self):
        return list(set(d.type for d in self.documents))
        
    labels = classes

    def _get_lsa(self):
        return self._lsa
    def _set_lsa(self, v=None):
        self._lsa = v
        self._update()
        
    lsa = property(_get_lsa, _set_lsa)

    def _get_weight(self):
        return self._weight
    def _set_weight(self, w):
        self._weight = w
        self._update() # Clear the cache.
        
    weight = property(_get_weight, _set_weight)

    @classmethod
    def load(cls, path):
        """ Loads the model from a pickle file created with Model.save().
        """
        return cPickle.load(open(path))