The thing that I completely don't understand is that when I run this script/model over and over again - I get completely different results each time - (sometimes 100% accuracy, sometimes 50% and sometimes 0%...) so it seems that the results are just random. Could someone explain to me why? (Note: I get random output with the same data! Test/train data are not shuffled/randomized)
I tried to simplify my model very much, including the dataset, so that it could be inserted here. The goal is to predict the outcome of a sports match from various statistical data. The outcome y should be either 1 (favorit wins), 0 (draw) or 2 (underdog wins).
import numpy as np
import tensorflow as tf
from keras.models import Sequential
from keras.layers import Dense
data = np.array ([
[2.0, 1.7, 4.82, 3.77, 23.0, 19.0, 21.0, 28.0, 3.0, 1.0, 23.0, 19.0, 21.0, 28.0, 3.0, 1.0, 16.0, 25.0, 14.0, 23.0, 23.0, 28.0, 23.0, 28.0, 28.0, 27.0, 17.0, 24.0, 29.0, 1.0, 10.0, 0.0, 29.0, 29.0, 0.0, 16.0, 12.0, 8.0, 8.0, 12.0, 5.0, 11.0, 8.0, 5.0, 11.0, 12.0, 11.0, 9.0, 11.0],
[1.0, 1.83, 4.73, 3.31, 8.0, 3.0, 3.0, 8.0, 12.0, 13.0, 8.0, 3.0, 3.0, 8.0, 12.0, 13.0, 10.0, 13.0, 12.0, 10.0, 23.0, 21.0, 23.0, 28.0, 28.0, 26.0, 17.0, 24.0, 29.0, 28.0, 13.0, 24.0, 5.0, 5.0, 24.0, 10.0, 12.0, 6.0, 8.0, 12.0, 5.0, 11.0, 8.0, 5.0, 11.0, 12.0, 11.0, 9.0, 11.0],
[0.0, 2.19, 3.5, 3.15, 14.0, 9.0, 10.0, 15.0, 20.0, 20.0, 14.0, 9.0, 10.0, 15.0, 20.0, 20.0, 13.0, 18.0, 13.0, 16.0, 23.0, 25.0, 23.0, 28.0, 28.0, 27.0, 17.0, 24.0, 29.0, 0.0, 13.0, 27.0, 2.0, 2.0, 27.0, 13.0, 12.0, 7.0, 8.0, 12.0, 5.0, 11.0, 8.0, 5.0, 11.0, 12.0, 11.0, 9.0, 11.0],
[1.0, 1.84, 4.68, 3.3, 20.0, 22.0, 29.0, 4.0, 4.0, 28.0, 20.0, 22.0, 29.0, 4.0, 4.0, 28.0, 25.0, 4.0, 3.0, 1.0, 21.0, 25.0, 20.0, 24.0, 24.0, 0.0, 19.0, 24.0, 2.0, 26.0, 4.0, 8.0, 21.0, 21.0, 8.0, 25.0, 2.0, 4.0, 8.0, 2.0, 5.0, 2.0, 8.0, 5.0, 11.0, 1.0, 11.0, 9.0, 12.0],
[1.0, 1.51, 6.1, 4.3, 17.0, 19.0, 27.0, 3.0, 2.0, 26.0, 17.0, 19.0, 27.0, 3.0, 2.0, 26.0, 0.0, 20.0, 2.0, 8.0, 22.0, 2.0, 19.0, 24.0, 24.0, 1.0, 19.0, 24.0, 2.0, 28.0, 2.0, 14.0, 15.0, 15.0, 14.0, 0.0, 3.0, 6.0, 8.0, 2.0, 5.0, 3.0, 8.0, 5.0, 11.0, 1.0, 11.0, 9.0, 12.0],
[1.0, 1.3, 11.52, 4.95, 22.0, 28.0, 10.0, 16.0, 13.0, 4.0, 22.0, 28.0, 10.0, 16.0, 13.0, 4.0, 23.0, 16.0, 4.0, 29.0, 21.0, 23.0, 20.0, 24.0, 24.0, 0.0, 19.0, 24.0, 2.0, 26.0, 17.0, 7.0, 22.0, 22.0, 7.0, 23.0, 2.0, 3.0, 8.0, 1.0, 5.0, 2.0, 8.0, 5.0, 11.0, 1.0, 11.0, 9.0, 12.0],
[1.0, 1.82, 5.11, 3.18, 25.0, 25.0, 5.0, 16.0, 17.0, 9.0, 25.0, 25.0, 5.0, 16.0, 17.0, 9.0, 14.0, 22.0, 7.0, 22.0, 21.0, 11.0, 20.0, 25.0, 25.0, 0.0, 19.0, 24.0, 2.0, 22.0, 15.0, 27.0, 2.0, 2.0, 27.0, 14.0, 2.0, 10.0, 8.0, 1.0, 5.0, 2.0, 8.0, 5.0, 11.0, 1.0, 11.0, 9.0, 12.0],
[2.0, 2.54, 2.97, 3.04, 18.0, 21.0, 28.0, 4.0, 4.0, 27.0, 18.0, 21.0, 28.0, 4.0, 4.0, 27.0, 5.0, 19.0, 1.0, 16.0, 22.0, 8.0, 19.0, 23.0, 23.0, 1.0, 19.0, 24.0, 2.0, 0.0, 20.0, 18.0, 11.0, 11.0, 18.0, 5.0, 3.0, 8.0, 8.0, 2.0, 5.0, 3.0, 8.0, 5.0, 11.0, 1.0, 11.0, 9.0, 12.0],
[1.0, 1.62, 2.26, 0.0, 21.0, 27.0, 23.0, 15.0, 9.0, 10.0, 21.0, 27.0, 23.0, 15.0, 9.0, 10.0, 9.0, 8.0, 11.0, 12.0, 28.0, 16.0, 10.0, 10.0, 10.0, 9.0, 21.0, 26.0, 7.0, 15.0, 24.0, 22.0, 7.0, 7.0, 22.0, 9.0, 6.0, 12.0, 10.0, 6.0, 7.0, 7.0, 9.0, 4.0, 10.0, 1.0, 11.0, 9.0, 12.0],
[1.0, 1.53, 2.46, 0.0, 25.0, 0.0, 26.0, 18.0, 12.0, 13.0, 25.0, 0.0, 26.0, 18.0, 12.0, 13.0, 11.0, 4.0, 11.0, 16.0, 28.0, 16.0, 10.0, 9.0, 9.0, 8.0, 21.0, 26.0, 7.0, 16.0, 18.0, 24.0, 5.0, 5.0, 24.0, 11.0, 6.0, 1.0, 10.0, 6.0, 7.0, 7.0, 9.0, 4.0, 10.0, 1.0, 11.0, 9.0, 12.0],
])
trainSize = int(len(data) * 0.8)
testSize = len(data) - trainSize
trainData, testData = data[0:trainSize], data[trainSize:]
xTrain, yTrain = trainData.T[1:].T, trainData.T[0]
xTest, yTest = testData.T[1:].T, testData.T[0]
model = Sequential()
model.add (tf.keras.layers.Dense(3, activation="softmax"))
model.compile(
optimizer='adam',
loss = 'mse',
metrics=['accuracy'],
)
model.fit(xTrain, yTrain, epochs=200, verbose=0)
val_loss, val_acc = model.evaluate(xTest, yTest)
print (val_loss, val_acc)
1/1 [==============================] - 0s 181ms/step - loss: 0.6667 - accuracy: 1.0000
0.6666609048843384 1.0
1/1 [==============================] - 0s 186ms/step - loss: 0.6667 - accuracy: 0.0000e+00
0.6666666865348816 0.0
1/1 [==============================] - 0s 230ms/step - loss: 0.6667 - accuracy: 1.0000
0.6666666865348816 1.0
1/1 [==============================] - 0s 263ms/step - loss: 0.6632 - accuracy: 0.5000
0.6631913781166077 0.5
1/1 [==============================] - 0s 196ms/step - loss: 0.6667 - accuracy: 0.0000e+00
0.6666661500930786 0.0
1/1 [==============================] - 0s 157ms/step - loss: 0.6667 - accuracy: 1.0000
0.6666666865348816 1.0
1/1 [==============================] - 0s 217ms/step - loss: 0.6589 - accuracy: 0.5000
0.6589114665985107 0.5
