【无标题】

=lightgbm==

在这里插入
import sqlite3
import pandas as pd
import numpy as np
import lightgbm as lgb
from sklearn.model_selection import train_test_split
from sklearn.metrics import classification_report

# 连接 SQLite 数据库
coon = sqlite3.connect("E:\code1\data2.db")

# 从数据库中读取三种类别的数据
positive1 = pd.read_sql_query("""SELECT * FROM data_1127 WHERE "defectName-1" LIKE "%多锡%" LIMIT 10000;""", coon)
positive2 = pd.read_sql_query("""SELECT * FROM data_1127 WHERE "defectName-1" LIKE "%少锡%" LIMIT 10000;""", coon)
negative = pd.read_sql_query("""SELECT * FROM data_1127 WHERE "defectName-1" LIKE "" LIMIT 10000;""", coon)

# 打印数据的头部信息
print(positive1.head(10))
print(positive2.head(10))
print(negative.head(10))

# 定义数据列
xcol = ['productEn', 'partName', 'pcbLength', 'pcbWidth',
       'pcbThickness', 'panelInformation', 'outlineDimension', 'sym_name',
       'sym_x', 'sym_y', 'pin_number', 'pin_x', 'pin_y', 'padshape1',
       'padwidth', 'padhght', 'solderPaster', 'board_side', 'coating', 'metal',
       'frame', 'step_thickness', 'step_qty', 'complexity_index',
       'part_number', 'x_res', 'y_res', 'thickness', 'area_ratio',
       'aspect_ratio', 'area', 'shape', 'symbol_l', 'symbol_w', 'version',
       'all_max_x', 'sym_rotate', 'x_min', 'x_max', 'y_min', 'all_max_y',
       'y_max']
num_col=['pcbLength', 'pcbWidth',
       'pcbThickness',
       'sym_x', 'sym_y',  'pin_x', 'pin_y',
       'padwidth', 'padhght',   'step_qty', 'complexity_index',
        'x_res', 'y_res', 'thickness', 'area_ratio',
       'aspect_ratio', 'area', 'symbol_l', 'symbol_w',
       'all_max_x', 'sym_rotate', 'x_min', 'x_max', 'y_min', 'all_max_y',
       'y_max']
cat_col = ['productEn', 'partName', 'panelInformation', 'outlineDimension', 'sym_name',
        'pin_number', 'padshape1',
        'solderPaster', 'board_side', 'coating', 'metal',
       'frame', 'step_thickness',
        'shape',  'version']

# 数据预处理
X = pd.concat([positive1[xcol], positive2[xcol], negative[xcol]])
X["pcbLength"] = X["pcbLength"].str.strip(" mm")
X["pcbWidth"] = X["pcbWidth"].str.strip(" mm")
X["pcbThickness"] = X["pcbThickness"].str.strip(" mm")
X["pcbLength"][X["pcbLength"] == ""] = 0
for col in num_col:
    X[col][X[col] == ""] = 0
    X[col] = X[col].astype(float)
X = X[num_col]

# 设置标签（0: 无缺陷, 1: 多锡, 2: 少锡）
Y = np.concatenate((np.ones(positive1.shape[0]), 2*np.ones(positive2.shape[0]), np.zeros(negative.shape[0])))

# 划分训练集和测试集
X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.2, random_state=42)

# 创建 LightGBM 模型
model = lgb.LGBMClassifier(num_leaves=31, learning_rate=0.05, n_estimators=100)

# 训练模型
model.fit(X_train, Y_train)

# 进行预测
ypred = model.predict(X_test)

# 评估模型
print(classification_report(Y_test, ypred))

# 完成
print("训练完成")