Pythonで機械学習を行っているのですが、エラーがでてしまい困っています

Question

前提・実現したいこと

機械学習で為替の予測をしているのですが、特徴量であるADXRを用いて学習をさせようとしたら、エラーメッセージがでてしまい先に進めることができません。

発生している問題・エラーメッセージ

ValueError                                Traceback (most recent call last)
<ipython-input-19-82ba81b7b0d8> in <module>
    135 #訓練用の説明変数と正解クラス、評価用の説明変数と正解クラスに分割
    136 for i in range(10):
--> 137     X_train,X_test,Y_train,Y_test = train_test_split(df_xy[["ADXR"]],df_xy["前日比_classified"],train_size=0.8)
    138 
    139     clf = RandomForestClassifier(max_depth=30,n_estimators=50)

~\miniconda3\lib\site-packages\sklearn\model_selection\_split.py in train_test_split(*arrays, **options)
   2129     n_samples = _num_samples(arrays[0])
   2130     n_train, n_test = _validate_shuffle_split(n_samples, test_size, train_size,
-> 2131                                               default_test_size=0.25)
   2132 
   2133     if shuffle is False:

~\miniconda3\lib\site-packages\sklearn\model_selection\_split.py in _validate_shuffle_split(n_samples, test_size, train_size, default_test_size)
   1812             'resulting train set will be empty. Adjust any of the '
   1813             'aforementioned parameters.'.format(n_samples, test_size,
-> 1814                                                 train_size)
   1815         )
   1816 

ValueError: With n_samples=0, test_size=None and train_size=0.8, the resulting train set will be empty. Adjust any of the aforementioned parameters.

該当のソースコード

import numpy as np 
import pandas as pd
import talib as ta
import matplotlib.pyplot as plt

df = pd.read_csv('USD_JPY 10-14.csv')


open = np.array(df["始値"])
close = np.array(df["終値"])
high = np.array(df["高値"])
low = np.array(df["安値"])

#特徴量を格納するデータフレーム
df_feature = pd.DataFrame(index=range(len(df)),columns=["ADXR","APO","AROON","AROONOSC","BOP","CCI","CMO","DX","MACDEXT","MACDFIX","MFI","MINUS_DI","MINUS_DM","MOM","PPO"])


#MACD
df_feature["MACD"],_,_ = ta.MACD(close,fastperiod = 12,slowperiod = 26,signalperiod = 9)

#ボリンジャーバンド
upper,middle,lower = ta.BBANDS(close,timeperiod = 5,nbdevup = 2,nbdevdn = 2,matype = 0)
df_feature["BBANDS+2σ"] = upper/close
df_feature["BBANDS-2σ"] = lower/close

#RSI
df_feature["RSI"] = ta.RSI(close, timeperiod=14)

#SMA5/current
df_feature["SMA5/current"]= ta.SMA(close, timeperiod=5) / close
df_feature["SMA20/current"]= ta.SMA(close, timeperiod=20) / close

#TEMA
df_feature["TEMA"]= ta.TEMA(close, timeperiod=30)

#WMA
df_feature["WMA"]= ta.WMA(close, timeperiod=30)

#MIDPOINT
df_feature["MIDPOINT"]= ta.MIDPOINT(close, timeperiod=14)

#ADX
df_feature["ADX"]= ta.ADX(high,low,close, timeperiod=14)

#DEMA
df_feature["DEMA"]= ta.DEMA(close, timeperiod=30)

#EMA
df_feature["EMA"]= ta.EMA(close, timeperiod=30)

#HT_TRENDLINE
df_feature["HT_TRENDLINE"]= ta.HT_TRENDLINE(close)

#KAMA
df_feature["KAMA"]= ta.KAMA(close, timeperiod=30)

#MA
df_feature["MA"]= ta.MA(close, timeperiod=30, matype=0)

#MIDPRICE
df_feature["MIDPRICE"]= ta.MIDPRICE(high, low, timeperiod=14)

#SAR
df_feature["SAR"]= ta.SAR(high, low, acceleration=0, maximum=0)

#ADXR
df_feature["ADXR"]= ta.ADXR(high, low, close, timeperiod=14)

#APO
df_feature["APO"]= ta.APO(close, fastperiod=12, slowperiod=26, matype=0)

#AROON
#df_feature["AROON"]= ta.AROON(high, low, timeperiod=14)

#AROONOSC
df_feature["AROONOSC"]= ta.AROONOSC(high, low, timeperiod=14)

#BOP
df_feature["BOP"]= ta.BOP(open, high, low, close)

#CCI
df_feature["CCI"]= ta.CCI(high, low, close, timeperiod=14)

#CMO
df_feature["CMO"]= ta.CMO(close, timeperiod=14)

#DX
df_feature["DX"]= ta.DX(high, low, close, timeperiod=14)

#MACDEXT
#df_feature["MACDEXT"]= ta.MACDEXT(close, fastperiod=12, fastmatype=0, slowperiod=26, slowmatype=0, signalperiod=9, signalmatype=0)

#MACDFIX
#df_feature["MACDFIX"]= ta.MACDFIX(close, signalperiod=9)

#MFI
#df_feature["MFI"]= ta.MFI(high, low, close, volume, timeperiod=14)


#教師データ
df["前日比_float"] = df["前日比%"].apply(lambda x: float(x.replace("%","")))

def classify(x):
    if x <= -0.2:
        return 0
    
    elif -0.2 < x < 0.2:
        return 1
    
    elif 0.2 <= x:
        return 2
    
df["前日比_classified"] = df["前日比_float"].apply(lambda x: classify(x))

df_y = df["前日比_classified"].shift()

df_xy = pd.concat([df_feature, df_y],axis = 1)

df_xy = df_xy.dropna(how = "any")


from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score
import optuna

list = []
list2 = []
list3 = []
sum = 0
sum2 = 0
sum3 = 0

#訓練用の説明変数と正解クラス、評価用の説明変数と正解クラスに分割 
for i in range(10):
    X_train,X_test,Y_train,Y_test = train_test_split(df_xy[["ADXR"]],df_xy["前日比_classified"],train_size=0.8)

    clf = RandomForestClassifier(max_depth=30,n_estimators=50)
    clf.fit(X_train, Y_train)

    print("訓練セットの精度"+str(i+1)+"回目: {:.3f}".format(clf.score(X_train, Y_train)))
    print("テストセットの精度"+str(i+1)+"回目: {:.3f}".format(clf.score(X_test, Y_test)))
    sum += round(clf.score(X_test, Y_test),3)
    list.insert(i,round(clf.score(X_test, Y_test),3)*100)
    
print(list)
print("合計：",sum)
ave = round(sum/10,3)
print("平均：",ave)

for j in range(10):
    X_train,X_test,Y_train,Y_test = train_test_split(df_xy[["ADXR"]],df_xy["前日比_classified"],train_size=0.8)

    clf = RandomForestClassifier(max_depth=30,n_estimators=50)
    clf.fit(X_train, Y_train)

    print("訓練セットの精度"+str(j+1)+"回目: {:.3f}".format(clf.score(X_train, Y_train)))
    print("テストセットの精度"+str(j+1)+"回目: {:.3f}".format(clf.score(X_test, Y_test)))
    sum2 += round(clf.score(X_test, Y_test),3)
    list2.insert(j,round(clf.score(X_test, Y_test),3)*100)

print(list2)
print("合計：",sum2)
ave2 = round(sum2/10,3)
print("平均：",ave2)

for k in range(10):
    X_train,X_test,Y_train,Y_test = train_test_split(df_xy[["ADXR"]],df_xy["前日比_classified"],train_size=0.8)

    clf = RandomForestClassifier(max_depth=30,n_estimators=50)
    clf.fit(X_train, Y_train)

    print("訓練セットの精度"+str(k+1)+"回目: {:.3f}".format(clf.score(X_train, Y_train)))
    print("テストセットの精度"+str(k+1)+"回目: {:.3f}".format(clf.score(X_test, Y_test)))
    sum3 += round(clf.score(X_test, Y_test),3)
    list3.insert(k,round(clf.score(X_test, Y_test),3)*100)

print(list3)
print("合計：",sum3)
ave3 = round(sum3/10,3)
print("平均：",ave3)


alist = (list,list2,list3)
fig, ax = plt.subplots()
bp = ax.boxplot(alist)
ax.set_xticklabels(['30:50','30:100','30:150'])

plt.title('2010-2014')
plt.xlabel('parameter(max_depth:estimators)')
plt.ylabel('accuracy')

plt.ylim([20,50])
plt.grid()

試したこと

ネットで調べたのですが詳しいことがわからず、困っています。
原因を教えてもらえないでしょうか？
よろしくお願いします。

補足情報（FW/ツールのバージョンなど）

ここにより詳細な情報を記載してください。

Answer 1

the resulting train set will be empty. Adjust any of the aforementioned parametersで調べてみましたらいくつかでてきました。

ファイルパスがおかしいというのが理由でしたので確認してみては如何でしょうか？

https://teratail.com/questions/228767
https://stackoverflow.com/questions/60043276/valueerror-with-n-samples-0-test-size-0-2-and-train-size-none-the-resulting-t

Comments

[Zarame]

回答ありがとうございます。
同じように調べて取り組みたいと思います。