LearnSL (version 1.0.0)

perceptron: Perceptron

Description

Binary classification algorithm that learns to separate two classes of data points by finding an optimal decision boundary (hyper plane) in the feature space.

Usage

perceptron(
  training_data,
  to_clasify,
  activation_method,
  max_iter,
  learning_rate,
  details = FALSE,
  waiting = TRUE
)

Value

List with the weights of the inputs.

Arguments

training_data

Data frame with already classified observations. Each column represents a parameter of the values. The last column contains the output, this means, the expected output when the other column values are inputs. Each row is a different observation. It works as training data.

to_clasify

Vector containing the parameters of the new value that we want to classify.

activation_method

Activation function to be used. It must be one of "step", "sine", "tangent", "linear", "relu", "gelu" or "swish".

max_iter

Maximum epoch during the training phase.

learning_rate

Value at which the perceptron will learn from previous epochs mistakes.

details

Boolean value. If it is set to "TRUE" multiple clarifications and explanations are printed along the code

waiting

If TRUE while details = TRUE. The code will stop in each "block" of code and wait for the user to press "enter" to continue.

Author

Víctor Amador Padilla,

Details

Functioning:

Step 1

Generate a random weight for each independent variable.

Step 2

Check if the weights classify correctly. If they do, go to step 4

Step 3

Adjust weights based on the error between the expected output and the real output. If max_iter is reached go to step 4. If not, go to step 2.

Step 4

Return the weights and use them to classify the new value

Examples

# example code
perceptron(db_per_or, c(1, 1, 1), "gelu", 1000, 0.1)
perceptron(db_per_and, c(0,0,1), "swish", 1000, 0.1, TRUE, FALSE)