机器学习进阶必备-10 个高效 Python 工具包完全指南

from cleanlab.classification import CleanLearning
from sklearn.linear_model import LogisticRegression

# 初始化清洗器
cl = CleanLearning(clf=LogisticRegression())
# 训练并识别问题数据
cl.fit(X_train, y_train)
# 查找标签问题
issues = cl.find_label_issues()

# 高级用法
# 获取置信度矩阵
confident_joint = cl.confident_joint
# 获取噪声标签的概率
label_quality_scores = cl.get_label_quality_scores()

from lazypredict.Supervised import LazyRegressor, LazyClassifier

# 回归任务
reg = LazyRegressor(verbose=0, ignore_warnings=True)
models_train, predictions_train = reg.fit(X_train, X_test, y_train, y_test)

# 分类任务
clf = LazyClassifier(verbose=0, ignore_warnings=True)
models_train, predictions_train = clf.fit(X_train, X_test, y_train, y_test)

# 查看模型性能比较
print(models_train)

import lux
import pandas as pd

# 基础使用
df = pd.read_csv("dataset.csv")
df.visualize() # 自动生成可视化建议

# 高级用法
# 指定感兴趣的变量
df.intent = ["column_A", "column_B"]
# 设置可视化偏好
df.set_intent_as_vis(["Correlation", "Distribution"])

from pyforest import *

# 使用时自动导入
df = pd.read_csv("data.csv")  # pandas自动导入
plt.plot([1, 2, 3])  # matplotlib自动导入

# 查看已导入的模块
active_imports()

from pivottablejs import pivot_ui

# 创建交互式数据透视表
pivot_ui(df)

# 自定义配置
pivot_ui(df, 
         rows=['category'], 
         cols=['year'],
         aggregatorName='Sum',
         vals=['value'])

import drawdata
import pandas as pd

# 创建交互式绘图界面
df = drawdata.get_data()

# 导出绘制的数据
df.to_csv('drawn_data.csv')

# 命令行使用
# black your_script.py
# 或在Python中使用
import black

# 格式化代码字符串
formatted_code = black.format_str(source_code, mode=black.FileMode())

# 格式化整个项目
# black .

# 检查模式（不实际修改文件）
# black --check .

from pycaret.classification import *

# 设置实验
exp = setup(data, target='target_column')

# 比较所有模型
best_model = compare_models()

# 创建模型
model = create_model('rf')  # 随机森林

# 调优模型
tuned_model = tune_model(model)

# 预测
predictions = predict_model(best_model, data=test_data)

# 保存模型
save_model(model, 'model_name')

import pytorch_lightning as pl
import torch.nn.functional as F

class MyModel(pl.LightningModule):
    def __init__(self):
        super().__init__()
        self.layer = nn.Linear(28*28, 10)
        
    def training_step(self, batch, batch_idx):
        x, y = batch
        y_hat = self(x)
        loss = F.cross_entropy(y_hat, y)
        self.log('train_loss', loss)
        return loss
        
    def configure_optimizers(self):
        return torch.optim.Adam(self.parameters(), lr=0.001)

# 训练模型
trainer = pl.Trainer(max_epochs=10, gpus=1)
trainer.fit(model, train_loader, val_loader)

import streamlit as st
import pandas as pd
import plotly.express as px

st.title("数据分析仪表板")

# 侧边栏配置
with st.sidebar:
    st.header("配置")
    option = st.selectbox("选择图表类型", ["散点图", "折线图", "柱状图"])

# 文件上传
uploaded_file = st.file_uploader("选择CSV文件")
if uploaded_file:
    df = pd.read_csv(uploaded_file)
    st.dataframe(df)
    
    # 数据统计
    st.write("数据统计摘要")
    st.write(df.describe())
    
    # 创建可视化
    if option == "散点图":
        fig = px.scatter(df, x='column1', y='column2')
    elif option == "折线图":
        fig = px.line(df, x='column1', y='column2')
    else:
        fig = px.bar(df, x='column1', y='column2')
        
    st.plotly_chart(fig)
    
    # 下载处理后的数据
    st.download_button(
        label="下载处理后的数据",
        data=df.to_csv(index=False),
        file_name='processed_data.csv',
        mime='text/csv'
    )