Hello Faranak Darabi ,

Absoutely, you can definitely use genetic programming to extract a Boolean function from a classification dataset! The idea is to set up genetic programming to search for logical expressions that can classify your data accurately.

Set up the problem: Treat each feature in your dataset as a Boolean variable (e.g., True/False, 0/1). If your data isn’t binary, you can binarise it by applying a threshold to turn continuous values into 0s and 1s.

Define the genetic programming setup: You’ll want to represent solutions as Boolean expressions, like trees or simple equations, where each node is an operator (AND, OR, NOT) and each leaf is a feature.

Set up a fitness function: This function measures how well each Boolean expression matches the dataset labels. For a classification dataset, the fitness score could be the percentage of correct predictions the expression makes on the dataset.

Run genetic programming:

Start with a random population of Boolean expressions.

Use crossover and mutation to combine and tweak expressions, keeping the ones that do better on your data.

Repeat this over multiple generations, with the aim of gradually evolving expressions that classify your data more accurately.

Extract the best expression: After enough generations, you’ll end up with the best Boolean function that the algorithm found, which should be a decent fit for your data.

If you’re working in Python, libraries like DEAP are great for this. You can set up the problem, define Boolean operators, and run the genetic algorithm to evolve a Boolean expression for your dataset. This approach lets the algorithm explore tons of possible combinations to find something that works well.

This process is definitely feasible and can be surprisingly effective for discovering interpretable logical rules from data, especially if you’re looking for a clear, rule-based model rather than a black-box one.

thanx

Dear Faranak Darabi ,

The performance of an information retrieval system is usually measured in terms of two different criteria, precision and recall. This way, the optimization of any of its components is a clear example of a multiobjective problem. However, although evolutionary programming have been widely applied in the information retrieval area, in all of these applications both criteria have been combined in a single scalar fitness function by means of a weighting scheme. In this paper, we deal with using of Genetic Programming in Information retrieval specially in optimizing of a Boolean query.

https://ceur-ws.org/Vol-152/paper3.pdf

Regards,

Shafagat

Dear Wasswa Shafik

thank you for your answer.

Dear Shafagat Mahmudova

thank you.