Skip to content

Kalvar/ios-KRSVM

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

16 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

About

KRSVM is implemented SVM of machine learning.

Podfile

platform :ios, '7.0'
pod "KRSVM", "~> 1.0.0"

How To Get Started

Import

#import "KRSVM.h"

Use SMO

KRSMO *smo = [[KRSVM sharedSVM] useSMO];

Use Linear Kernel Function

[smo.kernel useLinear];

Use RBF Kernel Function

The sigma that could be customized by your wishes, default value is 2.0, but some papers said 0.5, 1.0, 3.0, 5.0 that all can do training. Anyway, just try, nothing else.

[smo.kernel useRBF];
smo.kernel.sigma = 2.0f;

Use Hyperbolic Tangent (tanh) Kernel Function

The alpha of tangent could be customized by your wishes, default value is 1.0, but sometimes to be 2.0 is better, in this sample case we used 0.8 to do regression.

[smo.kernel useTanh];
smo.kernel.alpha = 0.8f;

Training Sample

smo.toleranceError = 0.001f;
smo.maxIteration   = 1000;
smo.constValue     = 1;
[smo.kernel useLinear];

// Setup the groups of classification and the target-value of group
[smo addGroupOfTarget:-1.0f];
[smo addGroupOfTarget:1.0f];

// Patterns
[smo addPatterns:@[@0.0f, @0.0f] target:-1.0f]; // x1
[smo addPatterns:@[@2.0f, @2.0f] target:-1.0f]; // x2
[smo addPatterns:@[@2.0f, @0.0f] target:1.0f];  // x3
[smo addPatterns:@[@3.0f, @0.0f] target:1.0f];  // x4

// One bias likes a net of neural network
[smo addBias:@0.0f];

// One input value by one weight that likes inputs & weights of neural network
[smo addWeights:@[@0.0f, @0.0f]];

[smo classifyWithPerIteration:^(NSInteger iteration, NSArray *weights, NSArray *biases) {
    //NSLog(@"%li Iteration weights : %@", iteration, weights);
    //NSLog(@"%li Iteration biases : %@", iteration, biases);
} completion:^(BOOL success, NSArray *weights, NSArray *biases, NSDictionary *groups, NSInteger totalIterations) {
    NSLog(@"===============================================");
    NSLog(@"%li Completion weights : %@", totalIterations, weights);
    NSLog(@"%li Completion biases : %@", totalIterations, biases);
    NSLog(@"%li Completion groups : %@", totalIterations, groups);
    NSLog(@"===============================================");
    // Verify & Directly Output
    [smo classifyPatterns:@[@[@2.0f, @2.0f], @[@3.0f, @0.0f]] completion:^(NSArray *weights, NSArray *biases, NSArray *results, NSDictionary *allGroups) {
        for( KRSVMPattern *pattern in results )
        {
            NSLog(@"direct classify to target %@", pattern.classifiedTarget);
        }
        NSLog(@"all groups : %@", allGroups);
    }];
}];

Version

V1.0.0

LICENSE

MIT.

About

Support Vector Machine in Machine Learning

Resources

License

Stars

Watchers

Forks

Packages

No packages published