Genetic Algorithm Framework
Project description
Genetic-Algorithm
Implementation of Genetic-Algorithm for solution finding (optimization)
Easy to use GA implementation. With parallel computing and info-prints. Simple and flexible for your optimal solution finding.
Usage
-
Get the code
-
Download the project and add to python module search path in your code
import sys sys.path.insert(0, '../path_to_GA_py_dir') # ./ => this folder # ../ => supfolder
-
Or pip install it (easier)
pip install Simple-Genetic-Algorithm
-
-
Import the class and helper function
from genetic_algorithm import GA, get_random
-
Create 2 functions and parameters
class Example_GA(GA): def calculate_fitness(self, kwargs, params): # return here the fitness (how good the solution is) # as bigger as better! # hint: if you have a loss, you should propably just return -1*loss # example for sklearn model: model = RandomForestRegressor(n_estimators=params["n_estimators"], ...) model = model.fit(kwargs["X_train"], kwargs["y_train"]) # predict y_pred = model.predict(kwargs["X_test"]) # calc mean absolute error loss y_true = np.array(kwargs["y_test"]) y_pred = np.array(y_pred) return - np.mean(np.abs(y_true - y_pred)) def get_random_value(self, param_key): if param_key == "name_of_parameter_1": return get_random(10, 1000) elif param_key == "name_of_parameter_2": return get_random(["something_1", "something_2", "something_3"]) elif param_key == "name_of_parameter_3": return get_random(0.0, 1.0) ... parameters = ["n_estimators", "criterion", "max_depth", "max_features", "bootstrap"]
-
Create and run genetic algorithm and pass the input, which will be used in the calculate_fitness function (in kwargs variable)
optimizer = Example_GA(generations=10, population_size=15, mutation_rate=0.3, list_of_params=parameters) optimizer.optimize(X_train=X_train, y_train=y_train, X_test=X_dev, y_test=y_dev)
Short explanation:
The kwargs are the inputs of optimize-method. These are the values which are needed to calculate the fitness. Maybe you can calculate the fitness without them, depending on what you are optimizing.
The list of parameters are the gene/the solution, so the parameters which are changed and optimized.
The get_random_value method return a random value for a given parameter, so that the solutions can be initialized and mutated.
Examples
License
Feel free to use it. Of course you don't have to name me in your code :)
-> It is a copy-left license
For all details see the license file.
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Hashes for Simple_Genetic_Algorithm-0.1.9.5.tar.gz
Algorithm | Hash digest | |
---|---|---|
SHA256 | ba64d1bd339b4eb8120791b7e30f47e2923b370ad55b5f24159baef15c092790 |
|
MD5 | 79220bf656bd15b3486b28b2ec0c6c0d |
|
BLAKE2b-256 | 8aea01d1e9cd51855cd81de991fff815d86d228a6e3879c1608286d16db40fcd |