#FGNW SEEDGENE SELECTION PCA

import pandas as pd
import numpy as np
from sklearn.decomposition import PCA
from sklearn.preprocessing import StandardScaler

# Assuming you already have your dataset as a pandas DataFrame
Data = pd.read_csv('/data/cyang12/FGNW/combined_one_ass_score.txt', sep = '\t', engine = 'python', index_col = 0)
#Data = pd.read_csv('/data/cyang12/FGNW/4_features.csv', sep = ',',index_col = 0)
normalized_data = (Data-Data.mean())/Data.std()
print(normalized_data)

import pandas as pd
import numpy as np
from sklearn.decomposition import PCA
from sklearn.preprocessing import StandardScaler

# Assuming you already have your dataset as a pandas DataFrame
# df = pd.read_csv('your_dataset.csv')

# Standardize the dataset
scaler = StandardScaler()
scaled_df = scaler.fit_transform(df)

# Perform PCA to reduce dimensions to (20000,1)
pca = PCA(n_components=1)
reduced_df = pca.fit_transform(scaled_df)

# The reduced DataFrame
reduced_df = pd.DataFrame(reduced_df, columns=['PC1'])

# To calculate the explained variance ratio
explained_variance_ratio = pca.explained_variance_ratio_
print(f"Explained Variance Ratio: {explained_variance_ratio}")

import matplotlib.pyplot as plt

def reconstruction_error(n_components, scaled_data):
    pca = PCA(n_components=n_components)
    transformed_data = pca.fit_transform(scaled_data)
    reconstructed_data = pca.inverse_transform(transformed_data)
    error = np.sum(np.square(scaled_data - reconstructed_data))
    return error

n_components_list = list(range(1, min(df.shape) + 1))
errors = [reconstruction_error(n, scaled_df) for n in n_components_list]

plt.plot(n_components_list, errors)
plt.xlabel('Number of Principal Components')
plt.ylabel('Reconstruction Error')
plt.title('PCA Reconstruction Error vs. Number of Principal Components')
plt.show()import matplotlib.pyplot as plt