model_visualization.py

from sklearn.manifold import TSNE 
from sklearn.decomposition import PCA 
import matplotlib as mpl
from matplotlib import rc
import matplotlib.pyplot as plt 
import pandas as pd 
from gensim.models import KeyedVectors 
# 그래프에서 마이너스 폰트 깨지는 문제에 대한 대처 
mpl.rc('font', family='AppleGothic') # 한글출력
mpl.rcParams['axes.unicode_minus'] = False 
#plt.rc('font', family='D2Coding') 

def show_tsne(): 
    tsne = TSNE(n_components=2) 
    X = tsne.fit_transform(X_show) 
    df = pd.DataFrame(X, index=vocab_show, columns=['x', 'y']) 
    fig = plt.figure() 
    fig.set_size_inches(30, 20) 
    ax = fig.add_subplot(1, 1, 1) 
    ax.scatter(df['x'], df['y'])

    for word, pos in df.iterrows(): 
        ax.annotate(word, pos, fontsize=10) 
        plt.xlabel("t-SNE feature 0") 
        plt.ylabel("t-SNE feature 1") 
        plt.show() 

def show_pca(): 
    # PCA 모델을 생성합니다 
    pca = PCA(n_components=2) 
    pca.fit(X_show) 
    # 처음 두 개의 주성분으로 숫자 데이터를 변환합니다 
    x_pca = pca.transform(X_show) 
    plt.figure(figsize=(30, 20)) 
    plt.xlim(x_pca[:, 0].min(), x_pca[:, 0].max()) 
    plt.ylim(x_pca[:, 1].min(), x_pca[:, 1].max()) 
    for i in range(len(X_show)): 
        plt.text(x_pca[i, 0], x_pca[i, 1], str(vocab_show[i]), fontdict={'weight': 'bold', 'size': 9}) 
    
    plt.xlabel("first label") 
    plt.ylabel("second label") 
    plt.show() 


model_name = './training_data/vector_clean_data_final_ver2_iter1000' 
model = KeyedVectors.load_word2vec_format(model_name) 
vocab = list(model.wv.vocab) 
X = model[vocab] 

# sz개의 단어에 대해서만 시각화 
sz = 1500 
X_show = X[:sz,:] 
vocab_show = vocab[:sz] 

#show_tsne() 
show_pca()