generated from LufsX/Experiment-LogReg
This commit is contained in:
+9
-15
@@ -46,19 +46,6 @@ def predict_logreg(inX, weights):
|
||||
probability = sigmoid(linear_output)
|
||||
return 1 if probability > 0.5 else 0
|
||||
|
||||
def preprocess_data(X, y):
|
||||
# 合并目标变量的类别
|
||||
y[y == 2] = 0
|
||||
y[y == 3] = 0
|
||||
|
||||
# 删除特征3和特征24~28
|
||||
X = np.delete(X, [3], axis=1)
|
||||
X = X[:, :23]
|
||||
|
||||
return X, y
|
||||
|
||||
|
||||
|
||||
def load_data(filepath):
|
||||
# 使用csv模块读取数据
|
||||
data = []
|
||||
@@ -66,8 +53,13 @@ def load_data(filepath):
|
||||
reader = csv.reader(file, delimiter=' ')
|
||||
for row in reader:
|
||||
if row: # 确保行不为空
|
||||
# 将非数值和空字符串转换为np.nan
|
||||
row = [float(x) if x.replace('.', '', 1).isdigit() else np.nan for x in row]
|
||||
# 将非数值和空字符串转换为np.nan,并确保所有行具有相同的列数
|
||||
row = [float(x) if x.replace('.', '', 1).isdigit() or x == '' else np.nan for x in row]
|
||||
# 确保所有行具有相同的列数,可以通过填充缺失的列或截断多余的列来实现
|
||||
while len(row) < 30: # 假设数据集应该有30列
|
||||
row.append(np.nan)
|
||||
while len(row) > 30:
|
||||
row.pop()
|
||||
data.append(row)
|
||||
|
||||
# 将列表转换为numpy数组
|
||||
@@ -97,6 +89,8 @@ def preprocess_data(X, y):
|
||||
|
||||
return X, y
|
||||
|
||||
# 确保其他函数(sigmoid, train_logreg, predict_logreg)保持不变
|
||||
|
||||
# 确保其他函数(sigmoid, train_logreg, predict_logreg)保持不变
|
||||
def main():
|
||||
# 更新文件路径以匹配实际位置
|
||||
|
||||
Reference in New Issue
Block a user