Ggthjy

Can an insurance company drop you after receiving a bill and refusing to pay?

How can I deliver in-universe written lore to players without it being dry exposition?

Find some digits of factorial 17

How much mayhem could I cause as a sentient fish?

Porting Linux to another platform requirements

Digits in an algebraic irrational number

Intern applicant asking for compensation equivalent to that of permanent employee

Why zero tolerance on nudity in space?

Writing a character who is going through a civilizing process without overdoing it?

A starship is travelling at 0.9c and collides with a small rock. Will it leave a clean hole through, or will more happen?

Would a National Army of mercenaries be a feasible idea?

Does paint affect EMI ability of enclosure?

How to limit sight distance to 1 km

Why isn't there a non-conducting core wire for high-frequency coil applications

Strange Sign on Lab Door

Advice for a new journal editor

Explain the objections to these measures against human trafficking

Injecting creativity into a cookbook

Why do no American passenger airlines still operate dedicated cargo flights?

What is the purpose of easy combat scenarios that don't need resource expenditure?

Why publish a research paper when a blog post or a lecture slide can have more citation count than a journal paper?

How can I get my players to come to the game session after agreeing to a date?

In Linux what happens if 1000 files in a directory are moved to another location while another 300 files were added to the source directory?

Avoiding morning and evening handshakes

Applying CNN for cross sectional data

Consistently inconsistent cross-validation results that are wildly different from original model accuracyUsing small CNN for de-noising on a full imageChoosing between prepadding and postpadding for variable length sequence data in a CNNResources for CNN example with KerasMy Keras CNN doesn't learnBatch data before feed into CNN networkkeras Sequential CNN for image data reshaping data issuesValue of loss and accuracy does not change over EpochsWrangling data for CNNUsing Image Data along with CSV file data for CNN model

0

$begingroup$

I have a data-set with around 1K features along with 200K observations and a target having 5 different categories. My target is amount categories containing how much a client can invest (e.g. 0-25K category 1, 25K-50K category 2 and so on). As for the features, I have quite rich info such as value of the total assets in a bank, total amount in a saving account, total amount in a current account, occupation code, if he/she has a loan or mortgage etc.

For my multi-class classification, I have tried GBM and MLP(Keras with Tensorflow backend) and got 0.69, 0.68 accuracy, respectively. As seen I could not beat GBM by MLP whose code is given below.

import time 



from datetime  import timedelta  from keras.wrappers.scikit_learn 

import KerasClassifier  from keras.layers import LeakyReLU



start_time = time.time()



df_results = pd.DataFrame()



np.random.seed(666)



def baseline_model():      

    model = Sequential()

    model.add(Dense(2000, input_dim=X_train.shape[1]))

    model.add(LeakyReLU())





for j in range(0,5):



    model.add(Dense(100))

    model.add(LeakyReLU())

    model.add(Dropout(0.3))



model.add(Dense(5, activation='softmax'))



opt = keras.optimizers.Adam(lr=0.00001)

model.compile(loss='categorical_crossentropy', optimizer=opt, metrics=['accuracy'])



return model





estimator = KerasClassifier(build_fn=baseline_model)

#print("Training...") 

estimator.fit(X_train, Y_train, epochs=50,batch_size=128,validation_split=0.2,verbose=2) 

predictions = estimator.predict(X_test) 

print(predictions) 

print(encoder.inverse_transform(predictions))

elapsed_time = time.time() - start_time

Now I want to try the same task with CNN (if it makes sense?) From the blog link it is stated that CNN should be used when there is a spatial relationship in data.

So are there examples where CNN has been successfully applied to non-image, non-text and non-time series data? Could you guide me how to prepare the input data?

python neural-network deep-learning keras cnn

share|improve this question

edited 15 hours ago

mlee_jordan

asked 15 hours ago

mlee_jordanmlee_jordan

1264

$endgroup$

add a comment | 

0

$begingroup$

I have a data-set with around 1K features along with 200K observations and a target having 5 different categories. My target is amount categories containing how much a client can invest (e.g. 0-25K category 1, 25K-50K category 2 and so on). As for the features, I have quite rich info such as value of the total assets in a bank, total amount in a saving account, total amount in a current account, occupation code, if he/she has a loan or mortgage etc.

For my multi-class classification, I have tried GBM and MLP(Keras with Tensorflow backend) and got 0.69, 0.68 accuracy, respectively. As seen I could not beat GBM by MLP whose code is given below.

import time 



from datetime  import timedelta  from keras.wrappers.scikit_learn 

import KerasClassifier  from keras.layers import LeakyReLU



start_time = time.time()



df_results = pd.DataFrame()



np.random.seed(666)



def baseline_model():      

    model = Sequential()

    model.add(Dense(2000, input_dim=X_train.shape[1]))

    model.add(LeakyReLU())





for j in range(0,5):



    model.add(Dense(100))

    model.add(LeakyReLU())

    model.add(Dropout(0.3))



model.add(Dense(5, activation='softmax'))



opt = keras.optimizers.Adam(lr=0.00001)

model.compile(loss='categorical_crossentropy', optimizer=opt, metrics=['accuracy'])



return model





estimator = KerasClassifier(build_fn=baseline_model)

#print("Training...") 

estimator.fit(X_train, Y_train, epochs=50,batch_size=128,validation_split=0.2,verbose=2) 

predictions = estimator.predict(X_test) 

print(predictions) 

print(encoder.inverse_transform(predictions))

elapsed_time = time.time() - start_time

Now I want to try the same task with CNN (if it makes sense?) From the blog link it is stated that CNN should be used when there is a spatial relationship in data.

So are there examples where CNN has been successfully applied to non-image, non-text and non-time series data? Could you guide me how to prepare the input data?

python neural-network deep-learning keras cnn

share|improve this question

edited 15 hours ago

mlee_jordan

asked 15 hours ago

mlee_jordanmlee_jordan

1264

$endgroup$

add a comment | 

$begingroup$

I have a data-set with around 1K features along with 200K observations and a target having 5 different categories. My target is amount categories containing how much a client can invest (e.g. 0-25K category 1, 25K-50K category 2 and so on). As for the features, I have quite rich info such as value of the total assets in a bank, total amount in a saving account, total amount in a current account, occupation code, if he/she has a loan or mortgage etc.

For my multi-class classification, I have tried GBM and MLP(Keras with Tensorflow backend) and got 0.69, 0.68 accuracy, respectively. As seen I could not beat GBM by MLP whose code is given below.

import time 



from datetime  import timedelta  from keras.wrappers.scikit_learn 

import KerasClassifier  from keras.layers import LeakyReLU



start_time = time.time()



df_results = pd.DataFrame()



np.random.seed(666)



def baseline_model():      

    model = Sequential()

    model.add(Dense(2000, input_dim=X_train.shape[1]))

    model.add(LeakyReLU())





for j in range(0,5):



    model.add(Dense(100))

    model.add(LeakyReLU())

    model.add(Dropout(0.3))



model.add(Dense(5, activation='softmax'))



opt = keras.optimizers.Adam(lr=0.00001)

model.compile(loss='categorical_crossentropy', optimizer=opt, metrics=['accuracy'])



return model





estimator = KerasClassifier(build_fn=baseline_model)

#print("Training...") 

estimator.fit(X_train, Y_train, epochs=50,batch_size=128,validation_split=0.2,verbose=2) 

predictions = estimator.predict(X_test) 

print(predictions) 

print(encoder.inverse_transform(predictions))

elapsed_time = time.time() - start_time

Now I want to try the same task with CNN (if it makes sense?) From the blog link it is stated that CNN should be used when there is a spatial relationship in data.

So are there examples where CNN has been successfully applied to non-image, non-text and non-time series data? Could you guide me how to prepare the input data?

python neural-network deep-learning keras cnn

share|improve this question

edited 15 hours ago

mlee_jordan

asked 15 hours ago

mlee_jordanmlee_jordan

1264

$endgroup$

import time 



from datetime  import timedelta  from keras.wrappers.scikit_learn 

import KerasClassifier  from keras.layers import LeakyReLU



start_time = time.time()



df_results = pd.DataFrame()



np.random.seed(666)



def baseline_model():      

    model = Sequential()

    model.add(Dense(2000, input_dim=X_train.shape[1]))

    model.add(LeakyReLU())





for j in range(0,5):



    model.add(Dense(100))

    model.add(LeakyReLU())

    model.add(Dropout(0.3))



model.add(Dense(5, activation='softmax'))



opt = keras.optimizers.Adam(lr=0.00001)

model.compile(loss='categorical_crossentropy', optimizer=opt, metrics=['accuracy'])



return model





estimator = KerasClassifier(build_fn=baseline_model)

#print("Training...") 

estimator.fit(X_train, Y_train, epochs=50,batch_size=128,validation_split=0.2,verbose=2) 

predictions = estimator.predict(X_test) 

print(predictions) 

print(encoder.inverse_transform(predictions))

elapsed_time = time.time() - start_time

share|improve this question

edited 15 hours ago

mlee_jordan

asked 15 hours ago

mlee_jordanmlee_jordan

1264

share|improve this question

edited 15 hours ago

mlee_jordan

asked 15 hours ago

mlee_jordanmlee_jordan

1264

asked 15 hours ago

mlee_jordanmlee_jordan

1264

asked 15 hours ago

mlee_jordanmlee_jordan

1264