ifaqih/ifnaivebayes

Naive Bayes library for PHP

1.0.0 2023-02-04 06:47 UTC

Requires

• php: >= 8.0

MIT

• ifaqih

README

Naive Bayes classifier (NBC) is a machine learning method that utilizes probability and statistical calculations proposed by British scientist Thomas Bayes, which predicts future probabilities based on past experience.

Installation

With Composer:

composer require ifaqih/ifnaivebayes

Use The Library

use IFaqih\AIMethods\NaiveBayes;

Set Data Method

Method to set the data to be processed.

class::set_data()

• Type: static
• Parameter data type: array $data
• Return data type: void

Prototype:

$data = [
    "attribute name"   =>  value,
    "attribute name"   =>  value,
    "attribute name"   =>  value
]

info:

• label: something that represents the group name
• attribute: parameters to be calculated in the classification process

Set Training Data Method

A method for organizing past classification data that will be used as the basis for future classifications.

class::set_training_data()

• Type: static
• Parameter data type: array|object $data, int $type_training_data = 0, ?string $key_label = null
• Return data type: void

Prototype 1:

$data = [
    [
        "attribute name"   =>  value,
        "attribute name"   =>  value,
        "attribute name"   =>  value,
        "label"            =>  value
    ]
];

$type_training_data = 0;
or
$type_training_data = NB_BY_ROWS;

Prototype 2:

$data = [
    "label" =>  [
        [
            "attribute name"   =>  value,
            "attribute name"   =>  value,
            "attribute name"   =>  value
        ]
    ]
];

$type_training_data = 1;
or
$type_training_data = NB_BY_GROUP_LABEL;

Prototype 3:

Set training data based on number of attributes and number of labels. If using this prototype does not require data input (in the set_data() method) and will skip the attribute calculation stage, because it is assumed to have performed attribute calculations manually.

$data = [
    "label" =>  [
        "attributes"    =>  [
            "attribute name"   =>  number of attributes on the label,
            "attribute name"   =>  number of attributes on the label,
            "attribute name"   =>  number of attributes on the label
        ],
        "total"         =>  number of labels
    ]
];

$type_training_data = 1;
or
$type_training_data = NB_BY_NUMBER;

info:

• label: something that represents the group name
• attribute name: parameters to be calculated in the classification process
• "$key_label" parameter is not needed when using the second prototype

Process Method

The method that will perform the classification calculation.

class::process()

• Type: static
• Parameter data type: ?array $data = null, array|object|null $training_data = null, int $type_training_data = 0, ?string $key_label = null, bool $boolean_returned = true, bool $clear_after = false
• Return data type: string|bool|null

info: do not include parameters if already set data using 'set' method, otherwise this parameter data will overwrite data which have been set using 'set' method

Get Result Method

Method to get the result of the classification process.

class::get_result()

• Type: static
• Parameter data type: not needed
• Return data type: string|int|float|null

Get Result Point Method

Method to get the points result from the classification process.

class::get_result_point()

• Type: static
• Parameter data type: ?string $label = null
• Return data type: object|float|int|null

info: parameter in this method is used to get the calculation result points which refers to the requested label and will be returned in float data type, if parameter is not filled or contains null it will return all calculation result points in object data type

Clear Method

This method will delete all data stored statically on the class.

class::clear()

• Type: static
• Parameter data type: not needed
• Return data type: void

Test Results

Number of training data = 100000

Example Prototype Training Data 1:

array(5) {
  ["result"]=>
  string(2) "TI"
  ["point"]=>
  object(stdClass)(3) {
    ["TI"]=>
    float(0.0020101198329071602)
    ["TM"]=>
    float(0.00048783922881731547)
    ["TP"]=>
    float(0.0011221592061979964)
  }
  ["start_at"]=>
  float(1666072755.080761)
  ["end_at"]=>
  float(1666072755.145308)
  ["execution_time"]=>
  float(0.06454706192016602)
}


Example Prototype Training Data 2:

array(5) {
  ["result"]=>
  string(2) "TI"
  ["point"]=>
  object(stdClass)(3) {
    ["TI"]=>
    float(0.0020101198329071602)
    ["TM"]=>
    float(0.00048783922881731547)
    ["TP"]=>
    float(0.0011221592061979964)
  }
  ["start_at"]=>
  float(1666072755.27086)
  ["end_at"]=>
  float(1666072755.291312)
  ["execution_time"]=>
  float(0.020452022552490234)
}


Example Prototype Training Data 3:

array(5) {
  ["result"]=>
  string(2) "TI"
  ["point"]=>
  object(stdClass)(3) {
    ["TI"]=>
    float(0.0020101198329071602)
    ["TM"]=>
    float(0.00048783922881731547)
    ["TP"]=>
    float(0.0011221592061979964)
  }
  ["start_at"]=>
  float(1666072755.291341)
  ["end_at"]=>
  float(1666072755.29135)
  ["execution_time"]=>
  float(8.821487426757812E-6)
}