Skip to content
This repository

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
tree: 5291e63736
Fetching contributors…

Cannot retrieve contributors at this time

file 298 lines (249 sloc) 8.985 kb
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298
//
// 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>


@implementation BKBayesianClassifier

@synthesize pools;

- (id)init
{
    self = [super init];
    if (self) {
        corpus = [[BKBayesianDataPool alloc] initWithName:@"__Corpus__"];
        pools = [[NSMutableDictionary alloc] init];
        dirty = YES;
        
        _tokenizer = [[BKTokenizer alloc] init];
    }
    return self;
}

- (void)dealloc
{
    [corpus release];
    [pools release];
    [super dealloc];
}

#pragma mark -
#pragma mark Serialization Methods
- (void)encodeWithCoder:(NSCoder*)coder
{
    [coder encodeObject:corpus forKey:@"Corpus"];
    [coder encodeObject:pools forKey:@"Pools"];
}

- (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;
}

- (id)initWithContentsOfFile:(NSString*)path
{
    self = [[NSKeyedUnarchiver unarchiveObjectWithFile:path] retain];
    if (self) {
    }
    return self;
}

- (BKBayesianClassifier*)classifierWithContentsOfFile:(NSString*)path
{
    return [[[BKBayesianClassifier alloc] initWithContentsOfFile:path] autorelease];
}

- (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)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];
        }
    }
}
- (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, 1);
        
        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 / 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 (int 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;
}

#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 Classification 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];
    
    NSMutableDictionary *result = [NSMutableDictionary dictionaryWithCapacity:42];
    for (NSString *poolName in pools) {
        BKBayesianDataPool *pool = [pools objectForKey:poolName];
        NSArray *tokensProbabilities = [pool probabilitiesForTokens:tokens];
        
        if ([tokensProbabilities count] > 0) {
            float probabilityCombined = [self robinsonCombinerOnProbabilities:tokensProbabilities];
            [result setObject:[NSNumber numberWithFloat:probabilityCombined] forKey:poolName];
        }
    }
    
    return result;
}

#pragma mark -
#pragma mark Classifier Informations
- (void)printInformations
{
    [self updatePoolsProbabilities];
    [corpus printInformations];
    for (NSString *poolName in pools) {
        [[pools objectForKey:poolName] printInformations];
    }
}

@end
Something went wrong with that request. Please try again.