KeyError(f"{not_found} not in index")

前提

python Chainerで電力使用量予測を行おうとしています。
参考サイト:https://qiita.com/mix_dvd/items/cf625bbb447797afa063#%E3%81%AF%E3%81%98%E3%82%81%E3%81%AB

実現したいこと

Chainerで電力使用量予測

発生している問題・エラーメッセージ

KeyError(f"{not_found} not in index")

該当のソースコード

import pandas as pd
import numpy as np

df_kw = pd.read_csv("juyo-2021.csv",encoding="utf-8_sig",skiprows=2)
df_kw["MW"] = df_kw["kW"] * 10

df_kw["DATETIME"] = df_kw.index.map(lambda : pd.to_datetime(df_kw.DATE[] + " " + df_kw.TIME[_]))

df_kw["MW"].plot(figsize=(15,4))

def read_temp(filename):
df_temp = pd.read_csv(filename,encoding="utf-8_sig",skiprows=4)
df_temp.columns = ["DATETIME","TEMP"]
df_temp.DATETIME = df_temp.DATETIME.map(lambda : pd.to_datetime())
return df_temp

df_temp_Fukui = read_temp("data-Fukui-2021.csv")
df_temp_Fukui.rename(columns = {'TEMP':'TEMP_Fukui'}, inplace=True)

df_temp_Fukui.TEMP_Fukui.plot(figsize=(15,4))

df = df_kw.copy()
df = df.merge(df_temp_Fukui,how="inner", on="DATETIME")

df["MONTH"] = df.DATETIME.map(lambda _: _.month)
df["WEEK"] = df.DATETIME.map(lambda _: _.weekday())
df["HOUR"] = df.DATETIME.map(lambda _: _.hour)

df = df.dropna()

cols = ["MONTH","WEEK","HOUR"]
for col in cols:
df = df.join(pd.get_dummies(df[col], prefix=col))

x_cols = ["TEMP_Fukui"] + df.columns.tolist()[14:]

X = df[x_cols]
y = df["MW"]

from sklearn import model_selection

X_train, X_test, y_train, y_test = model_selection.train_test_split(X, y, test_size=.2, random_state=42)

from sklearn.preprocessing import StandardScaler

scaler = StandardScaler()
scaler.fit(X_train)

X_train = scaler.transform(X_train)
X_test = scaler.transform(X_test)

from chainer import Chain, optimizers, Variable
import chainer.functions as F
import chainer.links as L

class MyChain(Chain):

def __init__(self, n_units=10):
    super(MyChain, self).__init__(
         l1=L.Linear(len(x_cols), n_units),
         l2=L.Linear(n_units, n_units),
         l3=L.Linear(n_units, 1))

def __call__(self, x_data, y_data):
    x = Variable(x_data.astype(np.float32).reshape(len(x_data),len(x_cols)))
    y = Variable(y_data.values.astype(np.float32).reshape(len(y_data),1))
    #print(x)

    pred = self.predict(x)
    #print(pred)
    return F.mean_squared_error(pred, y)

def predict(self, x):
    h1 = F.relu(self.l1(x))
    h2 = F.relu(self.l2(h1))
    h3 = self.l3(h2)
    return h3

def get_predata(self, x):
    return self.predict(Variable(x.astype(np.float32).reshape(len(x),1))).data

batchsize = 16
n_epoch = 200
n_units = 10

model = MyChain(n_units)
optimizer = optimizers.Adam()
optimizer.setup(model)

train_losses =[]
test_losses =[]

N = len(X_train)
print(N)
print(X_train.shape)
print(y_train.shape)
for epoch in range(1, n_epoch + 1):

perm = np.random.permutation(N)

sum_loss = 0
for i in range(0, N, batchsize):
    x_batch = X_train[perm[i:i + batchsize]]
    y_batch = y_train[perm[i:i + batchsize]] →エラー

    model.zerograds()
    loss = model(x_batch,y_batch)
    sum_loss += loss.data * batchsize
    loss.backward()
    optimizer.update()

average_loss = sum_loss / N
train_losses.append(average_loss)

loss = model(X_test,y_test)
test_losses.append(loss.data)

if epoch % 10 == 0:
    print("epoch: {}/{} train loss: {} test loss: {}".format(epoch, n_epoch, average_loss, loss.data))

y_pred = model.predict(X_test.astype(np.float32))

pd.DataFrame(np.c_[y_test, y_pred.array], columns=["act","pred"])[:100].reset_index(drop=True).plot(figsize=(15,4))
plt.show()