vectorizer.h

使用具有增量学习的监控式学习方法。包括几个不同的分类算法。

#ifndef VECTORIZER_H
#define VECTORIZER_H

#include "dictionary.h"
#include "languages.h"
#include "selector.h"
#include "tokenizer.h"
#include "vector.h"

typedef struct vectorizer_ vectorizer;

vectorizer *
vectorizer_new (tokenizer *tok);

vectorizer *
vectorizer_bool_create (tokenizer *tok);

vectorizer *
vectorizer_tf_create (tokenizer *tok);

vectorizer *
vectorizer_tfidf_create (tokenizer *tok);

vectorizer *
vectorizer_hadamard_create (tokenizer *tok);

int
vectorizer_set_normalizer (vectorizer *vec, void (*normalize) (vector *));

void
vectorizer_set_autobias (vectorizer *vec, int bool);

dict *
vectorizer_get_dictionary (vectorizer *vec);

void
vectorizer_set_dictionary (vectorizer *vec, dict *dt);

void
vectorizer_set_selector (vectorizer *vec, selector *sel);

vector *
vectorizer_normalize_vector (vectorizer *vec, vector *v, int new);

vector *
vectorizer_vectorize_text (vectorizer *vec, vector *v, const char *text,
			   int n, const char *lang, const char *charset);

int
vectorizer_save (vectorizer *vec, FILE *f);

vectorizer *
vectorizer_load (FILE *f, languages *langs);

#endif