Skip to content
Snippets Groups Projects
Commit 44f8097d authored by s2-gado's avatar s2-gado
Browse files

Upload New File

parents
No related branches found
No related tags found
No related merge requests found
Hide whitespace changes
Inline Side-by-side
Showing
with 290 additions and 0 deletions
GA.ipynb 0 → 100644
+ 290
− 0
View file @ 44f8097d
%% Cell type:markdown id: tags:
## Genetic Algorithm
%% Cell type:code id: tags:
``` python
import pandas as pd
import numpy as np
import random
from scipy.stats import pearsonr
from deap import base, creator, tools, algorithms
from sklearn.ensemble import RandomForestClassifier
from sklearn.tree import DecisionTreeClassifier
from sklearn.model_selection import cross_val_score
from sklearn.model_selection import RandomizedSearchCV
from sklearn.base import BaseEstimator
from sklearn.model_selection import train_test_split
import matplotlib.pyplot as plt
```
%% Cell type:code id: tags:
``` python
a_df = pd.read_csv('alzheimers_disease_data.csv')
df =a_df.drop(['DoctorInCharge', 'PatientID'], axis=1, inplace=True)
X = a_df.drop('Diagnosis', axis = 1)
y = a_df['Diagnosis']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)
```
%% Cell type:code id: tags:
``` python
#defining the fitness function
def fitness_function(individual):
selected_indices = [i for i, val in enumerate(individual) if val == 1]
if len(selected_indices) == 0:
return -1.0, # Penalize empty selection
correlation = []
# Get selected features
selected_features = X_train.iloc[:, selected_indices]
#from selected features calculate correlation using pearson correlation
for feature in range(len(selected_indices)):
r, _= pearsonr((selected_features.iloc[:, feature].values), y_train.values)
correlation.append(abs(r))
mean_correlation = np.mean(correlation)
model = DecisionTreeClassifier(random_state= 42)
model.fit(selected_features, y_train)
scores = cross_val_score(model, selected_features, y_train, cv=3, scoring='accuracy')
mean_accuracy = np.mean(scores)
#fitness_score = mean_correlation , mean_accuracy
return mean_correlation, mean_accuracy
```
%% Cell type:markdown id: tags:
### Using DEAP to create Genetic Algorithm
%% Cell type:code id: tags:
``` python
#setting up the GA
creator.create("FitnessMax", base.Fitness, weights=(1.0,1.0)) # Maximize fitness
creator.create("Individual", list, fitness=creator.FitnessMax)
toolbox = base.Toolbox()
toolbox.register("attr_bool", random.randint, 0, 1) # Binary encoding (0 = ignore, 1 = select)
toolbox.register("individual", tools.initRepeat, creator.Individual, toolbox.attr_bool, n=len(X.columns))
toolbox.register("population", tools.initRepeat, list, toolbox.individual)
#Operators
toolbox.register("mate", tools.cxTwoPoint) # Crossover
toolbox.register("mutate", tools.mutFlipBit, indpb=0.05) # Mutation probability
toolbox.register("select", tools.selNSGA2) # Multi-objective selection selNSGA2 (Non-dominated Sorting Genetic Algorithm II) is well-suited for multi-objective optimization.
toolbox.register("evaluate", fitness_function)
```
%% Output
C:\Users\sarah\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.11_qbz5n2kfra8p0\LocalCache\local-packages\Python311\site-packages\deap\creator.py:185: RuntimeWarning: A class named 'FitnessMax' has already been created and it will be overwritten. Consider deleting previous creation of that class or rename it.
warnings.warn("A class named '{0}' has already been created and it "
C:\Users\sarah\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.11_qbz5n2kfra8p0\LocalCache\local-packages\Python311\site-packages\deap\creator.py:185: RuntimeWarning: A class named 'Individual' has already been created and it will be overwritten. Consider deleting previous creation of that class or rename it.
warnings.warn("A class named '{0}' has already been created and it "
%% Cell type:code id: tags:
``` python
# --- Define Main GA Function ---
def main(pop_size, generations, cxpb, mutpb):
population = toolbox.population(n=pop_size)
# Evaluate initial population
fitnesses = list(map(toolbox.evaluate, population))
for ind, fit in zip(population, fitnesses):
ind.fitness.values = fit
for g in range(generations):
# Select the next generation
offspring = toolbox.select(population, len(population))
offspring = list(map(toolbox.clone, offspring))
# Apply crossover and mutationi
for child1, child2 in zip(offspring[::2], offspring[1::2]):
if random.random() < cxpb:
toolbox.mate(child1, child2)
del child1.fitness.values
del child2.fitness.values
for mutant in offspring:
if random.random() < mutpb:
toolbox.mutate(mutant)
del mutant.fitness.values
# Evaluate new offspring
invalid_ind = [ind for ind in offspring if not ind.fitness.valid]
fitnesses = list(map(toolbox.evaluate, invalid_ind))
for ind, fit in zip(invalid_ind, fitnesses):
ind.fitness.values = fit
# Update population
population = toolbox.select(offspring, k=len(population))
# Use sortNondominated to select the best individual(s)
fronts = tools.sortNondominated(population, len(population))
best_front = fronts[0] # Front 1 contains the best non-dominated solutions
# Find the individual with the highest correlation in the best front
#find individual with highest correclation
best_corr_individual = max(best_front, key=lambda ind: ind.fitness.values[0])
#fitness.values[0]: Corr score
# fitness.values[1]: Accuracy score
#loops through best_corr_individual and selects the corresponding feature names from X_train.columns
best_corr_features = [X_train.columns[i] for i in range(len(best_corr_individual)) if best_corr_individual[i] == 1]
best_corr_value = best_corr_individual.fitness.values[0] #corr of the best individual
# Find the individual with the highest accuracy in the best front
best_acc_individual = max(best_front, key=lambda ind: ind.fitness.values[1])
best_acc_features = [X_train.columns[i] for i in range(len(best_acc_individual)) if best_acc_individual[i] == 1]
best_acc_value = best_acc_individual.fitness.values[1]
combined_features = list(set(best_corr_features + best_acc_features))
# Print results in the desired format
#print(f"Correlation: {best_corr_value},\n Accuracy: {best_acc_value},\n Best Features: {combined_features}")
return best_corr_value, best_acc_value, combined_features
# # Get Best Individual
# best_individual = tools.selBest(population, k=1)[0]
# selected_features = [X.columns[i] for i in range(len(best_individual)) if best_individual[i] == 1]
# print("Selected Features using Pearson Correlation GA: \n", selected_features)
# return best_individual.fitness.values[0], selected_features
```
%% Cell type:code id: tags:
``` python
runs = 5
corr_values = []
acc_values = []
for i in range(runs):
best_corr_value, best_acc_value, selected_features = main(pop_size=300, generations=300, cxpb=0.9, mutpb=0.2)
print("=" * 50)
print(f"Run {i+1}:")
print(f"Best Correlation Score: {best_corr_value:.4f}")
print(f"Best Accuracy Score: {best_acc_value:.4f}")
print("\nSelected Features:\n", selected_features)
print("=" * 50, "\n")
corr_values.append(best_corr_value)
acc_values.append(best_acc_value)
```
%% Output
==================================================
Run 1:
Best Correlation Score: 0.2161
Best Accuracy Score: 0.9136
Selected Features:
['MemoryComplaints', 'Diabetes', 'FunctionalAssessment', 'ADL', 'PersonalityChanges', 'MMSE', 'BehavioralProblems']
==================================================
%% Cell type:code id: tags:
``` python
plt.scatter(corr_values, acc_values, c='blue', marker='o', label='Runs')
for i in range(runs):
plt.annotate(f'({corr_values[i]:.4f}, {acc_values[i]:.4f})',
(corr_values[i], acc_values[i]),
textcoords="offset points",
xytext=(5,5),
ha='right',
fontsize=10,
bbox=dict(boxstyle="round,pad=0.3", edgecolor='gray', facecolor='white'))
plt.xlabel('Correlation')
plt.ylabel('Accuracy')
plt.title('Trade-off between Correlation and Accuracy')
plt.grid(True)
plt.legend()
plt.show()
```
%% Output
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment