How to interpret predicted data from a keras model

I tried building a keras model to classify leaves from the leaf classification dataset on kaggle. After i compiled and trained the model, i used it to predict the name of the leaves in the testing images, but all i got is an array of integers. How can i exactly interpret those numbers in order to get the names of the leaves.

model = Sequential()

model.add(Dense(128, kernel_initializer="uniform", input_dim= 192, activation='relu'))
model.add(Dropout(0.25))
model.add(Dense(99, activation='softmax'))

model.compile(loss='categorical_crossentropy',
              optimizer='adam',
              metrics=['accuracy'])
model_history = model.fit(x=X_train,y=Y_train, epochs=500, batch_size= 32, validation_data=(X_val, Y_val), verbose=1)

predictions = model.predict(test_arr, batch_size=32, verbose=1)
computed_predictions = np.argmax(predictions, axis=1)
computed_predictions

array([51, 50,  1, 19, 14,  3,  3, 28, 84,  8, 43, 74, 75, 10, 52, 46, 45,
       73, 13, 71, 61, 68, 57, 77,  1, 70, 28, 15, 35, 70, 53, 74, 47, 50,
        4, 36, 14, 69, 36, 93,  8, 32,  8,  9, 71, 70, 38, 23, 26, 18, 17,
        5, 55, 94, 14, 86, 62, 33, 51, 54, 88, 56, 21, 59, 65, 11, 48,  5,
       13,  4, 54, 57, 29,  7, 31, 98, 92, 84, 25, 10, 61, 43, 85, 24,  1,
        2, 23, 83, 40, 22, 48, 90, 25, 21, 37, 56, 41, 95,  7, 49, 98, 77,
        3, 12, 31, 84, 53, 96, 64, 72, 93, 93, 67, 30,  8, 88, 60, 87,  6,
       57, 34, 34, 60, 17, 75, 27, 51, 73, 39, 23, 38,  2, 41, 61, 24, 97,
       29, 28, 68, 81, 42, 51, 86, 62, 60, 52, 95, 81, 42, 96, 95, 20, 59,
       35, 86,  1, 26, 38, 43, 75, 20, 60, 46, 79, 22, 79, 69, 87, 65, 97,
       75, 21, 29, 21, 11, 10, 58, 94, 27, 22, 15, 45, 89, 54, 43,  5, 23,
       94, 40, 49, 89, 72, 36, 11, 81, 95, 18, 91, 29, 64, 80,  6, 78, 45,
       28,  9, 78, 90, 44, 89, 92, 13,  2, 59,  0, 96, 70, 32, 29, 78, 91,
       55, 44, 38,  5, 60, 49, 58, 93, 67, 92, 88, 90, 79, 25, 37, 18,  0,
       76, 27, 70, 71, 44, 70, 32, 90, 30, 82, 34, 30, 82, 96, 48, 65, 57,
       64, 26, 53, 69, 73,  9,  3, 83, 26, 30, 63, 17, 22, 36, 63, 12, 78,
       36, 14, 27, 25, 67, 38, 20, 54, 76, 69, 67, 97, 80, 44, 92, 69, 23,
       21, 11, 51, 33, 77, 16, 11, 97,  1, 52, 39, 24, 52, 42, 17,  2, 73,
       96, 83, 88,  9, 63, 50, 16, 37, 87, 95,  3, 35, 83, 60, 59, 58,  0,
       79, 62, 38, 93, 68, 69, 46, 19, 46, 94, 18,  0, 33, 89, 40, 62, 48,
       42,  6, 31, 91, 73, 81, 12, 85, 26,  6, 79,  2, 22, 35, 43,  6, 80,
       78, 82,  5, 61, 37, 43, 33, 69, 56, 71, 45, 59, 42, 66, 86, 98, 83,
       90, 64, 82, 11, 79, 56, 56, 49, 48, 20, 74, 15, 33, 49, 89, 44,  7,
       35, 14, 55, 23, 34, 44, 32, 30, 36,  9, 72, 31, 61, 50, 82, 34, 28,
       22, 92, 72, 11, 19,  4, 87, 51, 80, 39, 84, 32, 66, 36, 41, 31, 80,
        4, 26, 68, 96, 20, 36, 34, 39, 56, 73, 76, 84,  7, 67, 37,  8, 95,
       85, 62, 10, 65, 41,  2, 83, 86, 41, 52,  3, 49, 47, 76, 52, 11, 26,
       88, 71, 45, 39, 66, 87, 75, 74,  7, 64, 65, 78, 63, 56, 21, 61, 88,
       62, 91, 59, 12, 74, 15, 85,  8, 66, 57, 83, 82, 72, 58, 96,  7, 67,
       66, 57, 66, 92, 35, 18,  9, 54, 91, 65, 19, 15, 10, 24, 71, 69, 48,
       39, 98, 16, 19, 45, 74,  6, 69, 42, 34, 71, 47, 85, 28, 85, 47, 25,
       27, 58, 68, 84, 97, 63, 97, 76, 81, 87, 77, 14,  0, 28, 41, 14, 12,
       33, 86, 46,  4,  4, 47, 30, 19, 58, 13, 77, 98,  5, 49, 72, 53, 32,
       77, 40, 68, 26, 92, 16, 81, 37, 14, 93, 80, 53, 46, 25, 50, 17, 37,
       93,  0, 20, 54, 10, 91, 40, 81, 53, 18, 27,  1, 12, 54, 73, 15],
      dtype=int64)

Topic neural keras neural-network

Category Data Science

1
answered at

Simple do backwards transformation of y_pred.

You label-encoded Y_train (i.e. every number 1-99 essentially represents a leaf) you you need to do is find mapping from leaf number and leaf name (thats in LabelEncoder) you would proceede something like this

le=LabelEncoder()
bla bal
Y_train=le.fit_transform()
predictions_test = le.inverse_transform(prediction_test)
1
answered at

you have 99 leaf species and 594 test records. Also the output is (594,). Here each value represents corresponding leaf species columns index. check the species data you would have received after One-HotEncoder/get_dummies. each value of "computed_predictions" corresponds to their respective column number. you can export species column_name to a list(of 99 labels) and get the labels using list index(i.e. computed_predictions respective value)

About

Geeks Mental is a community that publishes articles and tutorials about Web, Android, Data Science, new techniques and Linux security.