Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
tree: 6d5127ceef
Fetching contributors…

Cannot retrieve contributors at this time

357 lines (294 sloc) 10.542 kb
//
// BKBayesianClassifier.m
// Licensed under the terms of the BSD License, as specified below.
//
/*
Copyright (c) 2010, Samuel Mendes
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
* Neither the name of ᐱ nor the names of its
contributors may be used to endorse or promote products derived
from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#import <BayesianKit/BKBayesianClassifier.h>
#import <BayesianKit/BKTokenizer.h>
const NSString *BKCorpusDataPoolName = @"__BKCorpus__";
@implementation BKBayesianClassifier
@synthesize pools;
- (id)init
{
self = [super init];
if (self) {
corpus = [[BKBayesianDataPool alloc] initWithName:BKCorpusDataPoolName];
pools = [[NSMutableDictionary alloc] init];
dirty = YES;
_tokenizer = [[BKTokenizer alloc] init];
}
return self;
}
- (id)initWithContentsOfFile:(NSString*)path
{
self = [[NSKeyedUnarchiver unarchiveObjectWithFile:path] retain];
if (self) {
}
return self;
}
- (void)dealloc
{
[corpus release];
[pools release];
[super dealloc];
}
#pragma mark -
#pragma mark NSCoding Methods
- (id)initWithCoder:(NSCoder*)coder
{
self = [super init];
if (self) {
_tokenizer = [[BKTokenizer alloc] init];
dirty = YES;
corpus = [[coder decodeObjectForKey:@"Corpus"] retain];
pools = [[coder decodeObjectForKey:@"Pools"] retain];
}
return self;
}
- (void)encodeWithCoder:(NSCoder*)coder
{
[coder encodeObject:corpus forKey:@"Corpus"];
[coder encodeObject:pools forKey:@"Pools"];
}
#pragma mark -
#pragma mark Creation Methods
- (BKBayesianClassifier*)classifierWithContentsOfFile:(NSString*)path
{
return [[[BKBayesianClassifier alloc] initWithContentsOfFile:path] autorelease];
}
#pragma mark -
#pragma mark Saving Methods
- (BOOL)writeToFile:(NSString*)path
{
return [NSKeyedArchiver archiveRootObject:self toFile:path];
}
#pragma mark -
#pragma mark Pool Management
- (BKBayesianDataPool*)poolNamed:(NSString*)poolName
{
BKBayesianDataPool *pool;
pool = [pools objectForKey:poolName];
if (pool == nil) {
pool = [[[BKBayesianDataPool alloc] initWithName:poolName] autorelease];
[pools setObject:pool forKey:poolName];
dirty = YES;
}
return pool;
}
- (void)removePoolNamed:(NSString*)poolName
{
[pools removeObjectForKey:poolName];
dirty = YES;
}
- (void)mergePoolNamed:(NSString*)sourcePoolName withPoolNamed:(NSString*)destPoolName
{
BKBayesianDataPool *sourcePool = [pools objectForKey:sourcePoolName];
BKBayesianDataPool *destPool = [pools objectForKey:destPoolName];
if (!sourcePool || !destPool) return;
for (NSString *token in sourcePool) {
NSUInteger count = [sourcePool countForToken:token];
[destPool addCount:count forToken:token];
}
dirty = YES;
}
#pragma mark -
#pragma mark Probabilities
- (void)updatePoolsProbabilities
{
if (dirty) {
[self buildProbabilityCache];
dirty = NO;
}
}
- (void)buildProbabilityCache
{
for (NSString *poolName in pools) {
BKBayesianDataPool *pool = [pools objectForKey:poolName];
NSUInteger poolTotalCount = [pool tokensTotalCount];
NSUInteger deltaTotalCount = MAX([corpus tokensTotalCount] - poolTotalCount, 1u);
for (NSString *token in pool) {
NSUInteger corpusCount = [corpus countForToken:token];
NSUInteger poolCount = [pool countForToken:token];
NSUInteger deltaCount = corpusCount - poolCount;
float goodMetric;
if (poolTotalCount == 0) {
goodMetric = 1.f;
} else {
goodMetric = MIN(1.f, (float)deltaCount/(float)poolTotalCount);
}
float badMetric = MIN(1.f, (float)poolCount/(float)deltaTotalCount);
float f = badMetric / (goodMetric + badMetric);
if (fabs(f - 0.5f) >= 0.1) [pool setProbability:f forToken:token];
}
}
}
#pragma mark -
#pragma mark Combiners
- (float)robinsonCombinerOnProbabilities:(NSArray*)probabilities
{
NSUInteger length = [probabilities count];
float nth = 1.0f / (uint32_t)length;
float probs[length], inverseProbs[length];
NSUInteger idx = 0;
for (NSNumber *probability in probabilities) {
probs[idx] = [probability floatValue];
inverseProbs[idx] = 1.0f - [probability floatValue];
idx++;
}
float inverseProbsReduced = inverseProbs[0];
float probsReduced = probs[0];
for (NSUInteger i = 1; i < length; i++) {
inverseProbsReduced = inverseProbsReduced * inverseProbs[i];
probsReduced = probsReduced * probs[i];
}
float P = 1.0f - powf(inverseProbsReduced, nth);
float Q = 1.0f - powf(probsReduced, nth);
float S = (P - Q) / (P + Q);
return (1.0f + S) / 2.0f;
}
- (double)chi2PWithChi:(double)chi andDegreeOfFreedom:(int)df
{
double m = chi / 2.0;
double sum, term;
if ((df & 1) == 1) return -1.0;
sum = term = exp(-m);
for (int i = 1; i < (df / 2); i++) {
term *= m/i;
sum += term;
}
return MIN(sum, 1.0);
}
- (float)robinsonFisherCombinerOnProbabilities:(NSArray*)probabilities
{
NSUInteger length = [probabilities count];
uint32_t nth = (uint32_t)length;
double probs[length], inverseProbs[length];
NSUInteger idx = 0;
for (NSNumber *probability in probabilities) {
probs[idx] = [probability floatValue];
inverseProbs[idx] = 1.0f - [probability floatValue];
idx++;
}
double inverseProbsReduced = inverseProbs[0];
double probsReduced = probs[0];
for (NSUInteger i = 1; i < length; i++) {
inverseProbsReduced = inverseProbsReduced * inverseProbs[i];
probsReduced = probsReduced * probs[i];
}
double H = [self chi2PWithChi:(-2.0f * log(probsReduced)) andDegreeOfFreedom:(2 * nth)];
double S = [self chi2PWithChi:(-2.0f * log(inverseProbsReduced)) andDegreeOfFreedom:(2 * nth)];
return (1.0 + H - S) / 2.0;
}
#pragma mark -
#pragma mark Trainning Methods
- (void)trainWithFile:(NSString*)path forPoolNamed:(NSString*)poolName
{
NSError *error = nil;
NSString *content = [NSString stringWithContentsOfFile:path
encoding:NSUTF8StringEncoding
error:&error];
if (error) {
NSLog(@"Error - %@", [error localizedDescription]);
return;
}
[self trainWithString:content forPoolNamed:poolName];
}
- (void)trainWithString:(NSString*)trainString forPoolNamed:(NSString*)poolName
{
NSArray *tokens = [_tokenizer tokenizeString:trainString];
BKBayesianDataPool *pool = [self poolNamed:poolName];
[self trainWithTokens:tokens inPool:pool];
dirty = YES;
}
- (void)trainWithTokens:(NSArray*)tokens inPool:(BKBayesianDataPool*)pool
{
for (NSString *token in tokens) {
if (!token || [token isEqual:@""]) continue;
[pool increaseCountForToken:token];
[corpus increaseCountForToken:token];
}
}
#pragma mark -
#pragma mark Guessing Methods
- (NSDictionary*)guessWithFile:(NSString*)path
{
NSError *error = nil;
NSString *content = [NSString stringWithContentsOfFile:path
encoding:NSUTF8StringEncoding
error:&error];
if (error) {
NSLog(@"Error - %@", [error localizedDescription]);
return nil;
}
return [self guessWithString:content];
}
- (NSDictionary*)guessWithString:(NSString*)string
{
NSArray *tokens = [_tokenizer tokenizeString:string];
[self updatePoolsProbabilities];
return [self guessWithTokens:tokens];
}
- (NSDictionary*)guessWithTokens:(NSArray*)tokens
{
NSMutableDictionary *result = [NSMutableDictionary dictionaryWithCapacity:[pools count]];
for (NSString *poolName in pools) {
BKBayesianDataPool *pool = [pools objectForKey:poolName];
NSArray *tokensProbabilities = [pool probabilitiesForTokens:tokens];
if ([tokensProbabilities count] > 0) {
float probabilityCombined = [self robinsonFisherCombinerOnProbabilities:tokensProbabilities];
[result setObject:[NSNumber numberWithFloat:probabilityCombined]
forKey:poolName];
}
}
return result;
}
#pragma mark -
#pragma mark Sanitizing Methods
- (void)stripToLevel:(NSUInteger)level
{
for (NSString *token in [corpus allTokens]) {
NSUInteger count = [corpus countForToken:token];
if (count < level) {
for (NSString *poolName in pools) {
BKBayesianDataPool *pool = [pools objectForKey:poolName];
[pool removeToken:token];
}
[corpus removeToken:token];
}
}
}
#pragma mark -
#pragma mark Printing Methods
- (void)printInformations
{
[self updatePoolsProbabilities];
[corpus printInformations];
for (NSString *poolName in pools) {
[[pools objectForKey:poolName] printInformations];
}
}
@end
Jump to Line
Something went wrong with that request. Please try again.