Htydjtk

I'm working on a binary classification in Keras with a Tensorflow backend. No matter how much I tweak, I can't seem to get my model past a val_acc of 67%. Is there something I'm missing, or is this just simply as accurate as I can get with my data?

Link to the data I am using

My Code

Load and Balance dataset to a 1:1, and create validation data.

from sklearn.utils import resample



raw_data = pd.read_csv('Data.csv')



df_majority = raw_data[raw_data['RESULT']==0].iloc[1:-2,0:3].dropna()

df_minority = raw_data[raw_data['RESULT']==1].iloc[1:-2,0:3].dropna()



print(raw_data['RESULT'].value_counts())





df_majority_downsampled = resample(df_majority, 

                          replace=False,

                 n_samples=raw_data['RESULT'].value_counts()[1],

                             random_state=123)



# Combine minority class with downsampled majority class

df_downsampled = pd.concat([df_majority_downsampled,df_minority])



# Display new class counts

print(df_downsampled['RESULT'].value_counts())

print(numpy.unique(df_downsampled['RESULT']))



X = df_downsampled.iloc[1:-2,0:2].dropna()

Y = df_downsampled.iloc[1:-2,2:3].dropna()



X, XTest, Y, YTest = train_test_split(X, Y, test_size = 0.3, random_state = 0)



print(YTest['RESULT'].value_counts()) #Just a double check to make

Create Model

def create_model(activation):

    model = Sequential()



    model.add(Dense(128,activation=activation,input_dim=2))

    model.add(BatchNormalization())



    model.add(Dense(64,activation=activation))

    model.add(BatchNormalization())

    model.add(Dropout(0.2))



    model.add(Dense(32,activation=activation))

    model.add(BatchNormalization())

    model.add(Dropout(0.2))



    model.add(Dense(16,activation=activation))

    model.add(BatchNormalization())

    model.add(Dropout(0.2))



    model.add(Dense(8,activation=activation))

    model.add(BatchNormalization())

    model.add(Dropout(0.2))



    model.add(Dense(4,activation=activation))

    model.add(BatchNormalization())

    model.add(Dropout(0.2))



    model.add(Dense(2,activation=activation))

    model.add(BatchNormalization())

    model.add(Dropout(0.2))



    model.add(Dense(1,activation='sigmoid'))

    # load weights

    model.load_weights("weights.best.hdf5")

    model.compile(loss='binary_crossentropy', optimizer=Adam(lr=0.0001),metrics=['accuracy'])  



return model

model = create_model('relu')



filepath       ="weights.best.hdf5"

checkpoint     = ModelCheckpoint(filepath,monitor='val_acc',verbose=1,save_best_only=True,mode='max')

callbacks_list = [checkpoint]



history = model.fit(X,Y,epochs=2000,batch_size=32, shuffle = True,validation_data = (XTest,YTest), verbose = 0, 

     callbacks=callbacks_list)

Predict and get score

from sklearn.metrics import roc_auc_score



predict = model.predict_classes(X)

print(numpy.unique(predict))



#for index,val in enumerate(predict):

#print("Predicted: %s, actual: %s, for val %s"(val[0],Y.iloc[index].values,X.iloc[index].values))



predict = [val[0] for val in predict]

print("ras score: ",roc_auc_score(Y,predict))





predict = model.predict(numpy.array([0.0235,0.5]).reshape(-1,2))

print(predict[0][0])

Result using current model