How to Use GRSBA - grsba.com

The Generalized Robust Spatial Bifurcation Algorithm (GRSBA) is an advanced version of the RSBA designed to enhance flexibility and adaptability in data modeling. This guide will walk you through the steps to implement and use GRSBA in your projects.

Table of Contents

Introduction

The GRSBA leverages principles of spatial bifurcation to provide a robust framework capable of handling complex and noisy data. This algorithm accepts up to six input arrays and processes them to generate reliable and consistent predictions.

Requirements

Python 3.7 or higher: Ensure your environment has Python 3.7 or a newer version.
NumPy: For numerical operations and array handling.
TensorFlow: Depending on the implementation of the grbsa model.
GRSBA Library: Ensure you have the grbsa module from the repository.

*OBS: From 06/19/24, there was a rollback to version 1.1.9. So, the newest version 2.6.0 is equal to 1.1.9.

Example 1:

Step 1: Import Necessary Libraries

First, install the package.

pip install grsba

Certify the version installed is grsba=2.6.2

You need to import the required libraries. For this example, we will use NumPy for data manipulation.


import os
os.environ['TF_ENABLE_ONEDNN_OPTS'] = '0'

import numpy as np
import grsba as gr

Step 2: Define Normalization Function

Normalization is crucial for ensuring that all input data falls within the same range, typically [0, 1].

def normalize(data):
    return (data - np.min(data)) / (np.max(data) - np.min(data))

Step 3: Define GRSBA Function

Next, define the GRSBA function that processes the input data using a generalized exponent t.

def apply_grsba(inputs, Y, t=1.0):
    processed_inputs = [input_data ** t for input_data in inputs]
    processed_inputs = np.array(processed_inputs).T
    
    X = processed_inputs.shape[1]
    num_samples = processed_inputs.shape[0]
    
    if Y.shape[0] != num_samples:
        raise ValueError(f"Shape mismatch: y should have {num_samples} rows but has {Y.shape[0]}")
    
    model = gr.GRSBA(X=X, Y=Y, t=2.0)
    model.train(processed_inputs, Y)
    
    result = model.predict(processed_inputs)
    return result

Complete Example

Here is the complete implementation:


import os
os.environ['TF_ENABLE_ONEDNN_OPTS'] = '0'

import numpy as np
import grsba as gr

# Auxiliary functions and example usage
def normalize(data):
    return (data - np.min(data)) / (np.max(data) - np.min(data))

def apply_grsba(inputs, Y, t=1.0):
    processed_inputs = [input_data ** t for input_data in inputs]
    processed_inputs = np.array(processed_inputs).T
    
    X = processed_inputs.shape[1]
    num_samples = processed_inputs.shape[0]
    
    if Y.shape[0] != num_samples:
        raise ValueError(f"Shape mismatch: y should have {num_samples} rows but has {Y.shape[0]}")
    
    model = gr.GRSBA(X=X, Y=Y, t=2.0)
    model.train(processed_inputs, Y)
    
    result = model.predict(processed_inputs)
    return result

# Example input data arrays (values normalized between 0 and 1)
temperature = np.array([0.2, 0.4, 0.6, 0.8, 0.7, 1.0])
humidity = np.array([0.3, 0.5, 0.7, 0.9, 0.6, 1.0])
precipitation = np.array([0.1, 0.2, 0.4, 0.5, 0.3, 1.0])
wind_speed = np.array([0.4, 0.6, 0.8, 0.9, 0.5, 1.0])
solar_radiation = np.array([0.5, 0.6, 0.7, 0.8, 0.9, 1.0])
pressure = np.array([0.3, 0.4, 0.5, 0.6, 0.2, 1.0])

# Targets for the training data
y = np.array([
    [0.6, 0.7, 0.8, 0.9, 1.0, 0.5], 
    [0.5, 0.4, 0.3, 0.2, 0.9, 0.5], 
    [0.1, 0.3, 0.5, 0.7, 0.9, 0.5], 
    [0.2, 0.4, 0.6, 0.8, 1.0, 0.5], 
    [0.0, 0.9, 0.1, 0.8, 0.0, 0.5],
    [0.9, 0.0, 0.6, 0.9, 0.0, 0.5]
])

# Ensure y has the correct shape
y = y.T  # Transpose to match the input shape

inputs = [normalize(temperature), normalize(humidity), normalize(precipitation), normalize(wind_speed), normalize(solar_radiation), normalize(pressure)]

outputs = apply_grsba(inputs, y, t=0.5)

# Show results
print("Predicted Thermal Comfort Index by GRSBA:")
print(outputs)

By following these steps, you can implement and use the GRSBA in your own projects to handle complex and noisy data effectively.

See: How to interpret the output