Questions tagged [train]

training (or estimation) of statistical models or algorithms.

Filter by
Sorted by
Tagged with
0
votes
1answer
18 views

Should the distribution of my samples in my model be equal to to the original data set?

In my original data set I see a distribution of 70% belongs to label A and 30% belongs to label B. For my train, validation and test set I maintained the same ratio. However, I wonder whether this is ...
2
votes
1answer
74 views

Standardization on training only or also including testing data?

My question is very much related to this one: How to apply standardization/normalization to train- and testset if prediction is the goal? However, my testing data is not a single observation that I ...
0
votes
0answers
296 views

How to Split Time Series Data to train/test for RNN [duplicate]

Let's say I have a set of time series data with 32 time steps. My goal is to predict what the data value would be for the next time step, given data for 30 previous time steps. Would it be okay to ...
1
vote
0answers
199 views

I am trying to build a progressive auto encoder neural network and I am not sure how to discard old weights?

The goal of the network is simple, encode and decode images at a smaller scale and slowly increasing the network complexity, the input image size and its output quality. My current weights for my ...
1
vote
0answers
26 views

different between effect of episodes and time in DQN and where is the updating the experience replay

In DQN paper of DeepMind company, there are two loops one for episodes and one for running time in each step (one for training and one for different time-step of running). Am I right? Since, nothing ...
1
vote
1answer
286 views

Bootstrapping accuracy, f1-score?

I have typical train/test setting, with an ordinary dataset. As I am comparing performance of two approaches to a problem (namely churn prediction with AdaBoost and ...
1
vote
1answer
225 views

In which scenarios are the in-sample error and training error NOT the same?

In Elements of Statistical Learning, Chapter 7 (pages 228-229), the authors define the optimism of the training error rate as: $$ op\equiv Err_{in}-\overline{err} $$ With the training error $\...
4
votes
1answer
595 views

When to stop training of neural network when validation loss is still decreasing but gap with training loss is increasing?

During training of CNNs, I often come across this case for training and validation loss : X axis is epochs, Y axis is cross entropy loss. I would like to keep the "best model", meaning the one which ...
3
votes
1answer
45 views

Can a ConvNet see patterns that a human cannot?

I am training a ConvNet to detect different types of stripes in my images. As I am working on astronomical images, my pixel values are flux densities and therefore represent ground truth data. When I ...
1
vote
1answer
348 views

Interpretation of logistic regression with normalized features

With logistic regression, a one unit change in $X_1$ is associated with a $\beta_1$ change in the log odds of 'success' (alternatively, an $\exp(\beta_1)$-fold change in the odds), all else being ...
8
votes
2answers
4k views

Is there a way to incorporate new data into an already trained neural network without retraining on all my data in Keras?

I have already trained a neural network on my data. In the future, I will receive some more data. How can I incorporate this data into my model without rebuilding it from scratch?
4
votes
3answers
154 views

When is there a difference between a normal likelihood loss and a least squares loss?

My understanding is that if the errors follow a normal distribution, then using a maximum likelihood loss or a least squares loss to train a model amounts to the same thing. However, I am looking at ...
1
vote
1answer
267 views

train / validation / test split

I understand that you typically use three different data sets (train/validation/test) to acquire an unbiased estimate of the performance measurement, because the models are tuned to fit for the train ...
3
votes
1answer
155 views

When predicting (not training) using neural networks, why would we have to specify a number of epochs?

I'm looking at code from a Google course on how to use Tensorflow. When explaining how to specify the function for generating predictions from an already trained model, the function they define takes ...
2
votes
0answers
136 views

Reducing batch size after X epochs?

Batch size and the number of iterations are considered as a tradeoff. It has been observed in practice that when using a larger batch there is a significant degradation in the quality of the model, ...
1
vote
1answer
32 views

Is it OK to use the parameters for the lowest cost? [duplicate]

In a Neural Network training, the cost of the model changes throughout the training process when using gradient descent (or something analogous), this is the point of the algorithm. However the cost ...
0
votes
1answer
64 views

Is it recommended to train a SVM model with the same dataset used for pre-train an autoencoder?

I have a very limited dataset and have used 80% of it to pre-train an autoencoder. Now, I attached the enconder part to a SVM. In order to train the SVM, is ok to train it using the exactly same (80%)...
0
votes
1answer
39 views

Dataset requirement for Deep Learning

I am doing research on deep neural networks for prediction. I wanted to know what minimum size of dataset is required for training a deep network. Is there any limitation imposed on how much ...
2
votes
2answers
430 views

Difference between train/test and train/validate/test split?

I know this question has been asked here before, but after reading the answers I still dont get the difference. Consider for instance a lasso penalized linnear regression model.This model has a ...
1
vote
0answers
18 views

How should Training Data for Fully ConvNets look like?

I've been working with CNNs recently. For a new task, I need to predict objects in an image pixelwise. Fully ConvNets seem to be the way to go. I read the original paper (Long et al., 2014) and a few ...
6
votes
1answer
624 views

Why would somebody use a hash function for creating a test/train split instead of random seed?

I'm going through some ML training material from Google (I can't post a link because I'm getting the material through my company). In the part about how to extract data and split it into train and ...
2
votes
2answers
304 views

Imputing the mean value from the 'train set' into the 'test set'

I have looked at a couple questions and answers similar to this, the recommendation seems to be the imputation of mean values from the 'training set' into my 'test set'. However, what I am trying to ...
2
votes
0answers
22 views

Is there an official rule, or a generally accepted one, for how closely your validation data should match your training data?

I'm finding myself making a gut-feeling judgment more and more often on whether the validation of my model is "close enough" to the modeled results from my training data. I don't recall having ever ...
2
votes
0answers
463 views

Creating training and validation sets for churn model

I need to determine a statistically sound methodology for creating training and validation datasets for a churn model. Testing sets and model selection aren't a problem. The data spans 4 years of ...
0
votes
1answer
71 views

How to encourage certain activations during training of neural networks?

Is it possible to train neural networks such that certain activations are rewarded and some other activations are penalized? In other words, I would like the network to generate preferred values more ...
0
votes
1answer
746 views

What % is the best train/test split for Time Series Data?

What % is the best train/test split for Time Series Data? Do you think it is still 70/30 ? I am talking about leaving a whole continuous period (in the beginning or in the end) for TEST. I have ...
0
votes
1answer
46 views

Some weights in CNN remain constant

When training my CNN, I notice that after several SGD updates, some weights of the layers do not change any more. Is this a normal situation? Will all the weights of the network layers change during ...
8
votes
1answer
645 views

How can the AIC or BIC be used instead of the train/test split?

I've recently come across several "informal" sources that indicate that in some circumstances, if we use the AIC or BIC to train a time series model, we don't need to split the data into test and ...
1
vote
0answers
62 views

What is difference between training examples generated by continuous bag of words(CBOW) and skip-gram?

This is a simple question that is hard for me: Let's consider simple sentence A B C D and create training examples for skip-gram training (x, y) with number of ...
1
vote
1answer
519 views

Test and Training dataset correlation while Splitting the dataset

I want to split my main dataset in two part, training dataset and test dataset. In the past i read somewhere (which unfortunately i could not find exactly where was that), that when splitting my ...
0
votes
1answer
107 views

Why steep loss reduction is an indication of inadequate initial weight allocation in neural network training?

This article says At times, you might see that your loss drops steeply after a short period of training, before stabilizing. This is a strong indication that your initial weight allocation is ...
0
votes
1answer
44 views

Duplicates in feature matrix

I have several points which appear duplicates in the feature matrix (same values for the features). These points may have different values of the target variable. What is the appropriate way to handle ...
1
vote
2answers
531 views

Is there a formula for a recommended batch size depending on the size of the training dataset?

I'm still training my neural network for gender/age classification, and I'm currently experimenting with batch sizes along with everything else. As I've gathered, too small a batch size will lower ...
3
votes
0answers
321 views

How to not overlook rare but important features when preventing over-fitting in a decision tree?

I have a data set where some binary features divide the sample space roughly in half, whereas other features are much less frequent and occur only for 0.0001 - 0.01 of the sample space. However, those ...
0
votes
2answers
40 views

How to know the predicted values on training data

How to know the predicted values of the model on training data? I just see the standard metrics such as RMSE and R-squared after run train () function.
0
votes
1answer
14 views

What images to train on when discriminating against a certain class

I am building an image classifier that discriminates against a certain class. As a toy example, let's say the classifier checks if the image is a hotdog or not (1 or 0). My question is what images ...
0
votes
0answers
30 views

RSS and $R^2$ are not suitable for selecting the best model, why? [duplicate]

This is in continuation to question Are RSS and R^2 related to training error only? In section 6.1.3 Choosing the Optimal Model of An Introduction to Statistical ...
4
votes
1answer
3k views

Batch normalization: How to update gamma and beta during backpropagation training step?

The backpropagation step of batch normalization computes the derivative of gamma (let's call it dgamma) and the derivative of <...
0
votes
0answers
33 views

Should one scale the full dataset or the training set directly [duplicate]

I'm following a MOOC at the moment and they seem to suggest when scaling it's best to scale the training data first and then use the training set's parameters to scale the test set. I see no reason ...
2
votes
3answers
131 views

How to use residuals to choose the best model parameters and avoid over-fitting?

A common way to avoid an over-fitting is to train a model on one set and then to check the error on the validation set. If the out-of-sample error is much larger than the in-sample error than we have ...
3
votes
2answers
50 views

How to regularize if zero values of model parameters do not give a “simple” model?

I have a predictive model with a relative small number of model parameters (only 6). When I train the model on the training set and than validate the model on the validation set I have a strong ...
0
votes
1answer
333 views

Should we remove duplicates when training a SVM?

Let's say I have a set P of positive examples and a set N of negative examples. Prior to feeding this dataset to a SVM for training, should I remove duplicates in those sets? Intuitively, I don't ...
1
vote
0answers
30 views

Using binary feedback (good / bad) to train a model with real-valued output

I have a model (which I would initialize with "best guess" parameters) which produces a real-valued output (which I need to be real-valued - I can't switch to a different representation), but for ...
0
votes
1answer
132 views

Help understanding training stacked autoencoders

I've been learning about stacked autoencoders, but wasn't entirely sure how to train them. From what I understand, given layers $h_1,h_2,...,h_n$, we greedily train as follows ...
4
votes
2answers
1k views

Is it always possible to achieve perfect accuracy on a small dataset?

I have read many times that a good debugging step while building a machine learning model is to try to overfit your model to a very small subset of your data. [Here is one such instance][1]. ...
2
votes
1answer
324 views

Definition of Overfitting

Consider error of hypothesis h over training data errortrain(h) and error over entire distribution D of data errorD(h). Hypothesis h in H overfits training data if there exists an alternative ...
1
vote
1answer
452 views

Many values of HMM matrices, A and B, tend to zero

I'm experimenting with an HMM. I have a sequence of observations (10000) and the original matrices A,B and pi that generated those observations. There are 4 types of observations. What I am trying to ...
0
votes
1answer
40 views

Iterative swarm training of multiple models

I've had an idea of a training scheme for multiple machine learning models, and want to know if it makes sense or it already has a name. The idea is to train models kinda like a swarm mind (I was ...
2
votes
1answer
360 views

Amount of training data for classification accuracy

Is there an intuition or any relevant reading about the relationship between dimensionality of data, number of samples, model complexity and test accuracy of classification? E.g. for the simple cat/...
0
votes
2answers
284 views

Can I add data, that my neural network classified, to the training set, in order to improve it?

Let's assume the following: I successfully trained a neural network on a classification task, it performs well, also on unseen data. Now my idea is: If the neural network obtains new, unseen data and ...