Questions tagged [train]
training (or estimation) of statistical models or algorithms.
287
questions
0
votes
0answers
13 views
How to train a model using a daunting huge training dataset
I have a extremely huge dataset and I'm wondering me how could be the right way to set an experiment to use this data to train a model.
I understand that I can use data-reduction to, for instance, ...
0
votes
0answers
4 views
How to evolve an already trained model with new data
How to evolve a model when new data come out?
Suppose I have a model $M_1$ trained on data $D$.
After the model $M_1$ is trained, new data $T$ is available.
How to train a new model $M_2$ from $M_1$ ...
0
votes
1answer
36 views
Random Forest Regression and Overfitting
I did gridsearch with corss-validation on a trainingset to search for best hyperparameters for a Random Forest Regressor. And indeed the best parameterset gives good results in cross-validation (R^2 ~ ...
0
votes
0answers
11 views
How to compute TBATS forecast accuracy without specific test set?
I have used the TBATS model on my data and when I apply the forecast() function, it automatically forecasts two years in the future. I haven't specified any training set or testing set, so how do I ...
0
votes
1answer
22 views
Method for selecting training data and identify predictive genes in a survival model
I am performing a coxph survival analysis and want to split my data into a training and test set. During the training I would like to identify the "best" ...
0
votes
1answer
32 views
How is stratified sampling better than sampling equally from all classes while crossvalidating?
I can see that stratified sampling helps in maintaining the same class distribution in the training set as in the original dataset. However, my understanding is that ideally, the model should be ...
1
vote
2answers
35 views
Training a Neural Network on Music
I'm very picky about the music I like. I can't say I like a genre in particular because my tastes in music boil down to how a specific song sounds. For a while now, I've had a vision of creating a ...
0
votes
0answers
12 views
High precision, but low recall on training data. What could be wrong?
I am training RNN for timeseries binary classification. I observed that network has high precision, but low recall on both training and testing data. I tried multiple architectures, but same problem ...
1
vote
1answer
122 views
Why not use large ($n - [n^{0.75}]$) validation sets in machine learning training loop?
I am seeking feedback in the form of answers to a bold question. Please limit comments to requests for clarification of the question only.
There appears to be an influence on machine learning ...
2
votes
0answers
20 views
Uplift modeling for Train, validation, test data sets
I am wondering when I should tune hyper parameter when we build uplift model.
In a normal machine learning context, data will be split into train, validation, test.
And, we train the model with train ...
0
votes
0answers
20 views
training variational autoencoder - loss is 0 val loss is 0
I am training a VAE in Keras on the colab. The issue that I am encountering is that after the first epoch the loss is 0 and the validation loss is 0 (but the model didn't learn anything).
this is the ...
0
votes
0answers
36 views
Why does switching the architecture result in nans during training?
What are my options to figure out the casue of spontaneous nans/infs during training?
Basically I changed the model from a custom resnet to the vanilla resnet(and also later a vggnet varation) and ...
3
votes
2answers
42 views
R2 on out-sample data set
The conventional definition of $R^2$ is:
$R^2 = 1-SSE/SST$, where SSE denotes sum of squared errors and SST is total sum of squares ($n\times variance$, n being number of sample points in train set).
...
0
votes
0answers
23 views
Small sized training set and results varying based on cross-validation split
I need to try to build a classifier based on around 30 instances. The outcome can also be that the dataset is not large enough for this purpose, however I'm not sure on how can I justify this outcome ...
4
votes
1answer
70 views
Does it make sense to regularize the loss function for binary/multi-class classification?
When discussing linear regression it is well known that you can add regularization terms, such as,
$$\lambda \|w\|^2 \quad \text{(Tikhonov regularization)}$$
to the empirical error/loss function.
...
0
votes
0answers
16 views
Using a poset or directed graph as input for a neural network
I'm not sure if this is the right community to post this in but I would appreciate any help. As the title states, I'm trying to train a neural network using some unconventional input. I'm wondering if ...
0
votes
0answers
7 views
How to train sentiment analysis with neutral
I want to classify sentiment on blog and forum sites. Each post can be positive, negative or neutral. Most of the existing sentiment classification are for two classes, positive and negative. I am ...
1
vote
0answers
44 views
Unit Deviance on test set R
I am evaluating a set of predictions coming from different models against a set of actual values in a regression problem.
I do not want to use the mean squared error as my evaluation metric because my ...
0
votes
1answer
21 views
validation error and test error when limited data is available
In machine learning, to get an unbiased estimate of model performance, we split data 80:20 into train and test set. We use the training set for model training and model selection according to cross-...
1
vote
0answers
8 views
How easily train model with set of hyperparameters with different algorithms?
I have tried hyperprameter optimization and do get top 15 best combination of parameters using gs.cv_result but now I would like to train a model on these 15 combination of hyperprameters seperately....
0
votes
0answers
57 views
Neural network training converges for several epochs, then diverges badly
I have a VGG-like network that I have trained from scratch on a multi-class dataset of my own. The results suggested there were probably some data errors somewhere, so I thought I would train the same ...
1
vote
0answers
37 views
Predicting when and where a conflict is most likely to take place (glm/ train test set error)
I am having difficulty analysing this dataset (I am new to statistics and R). I am trying to model when a conflict is most likely to take place (before, during or after deforestation), and at what ...
10
votes
3answers
638 views
Is it in general helpful to add “external” datasets to the training dataset? [closed]
Several people have already asked "is more data helpful?":
What impact does increasing the training data have on the overall system accuracy?
Can increasing the amount of training data make ...
0
votes
0answers
28 views
How to evaluate xgb.cv and is it okay to use min(test_logloss) even if there is large gab between train_logloss?
Please migrate if this is not correct spot the post this question.
I have a dataset including 70K - 1s and 300K - 0s, a type of classifier. I am using Xgboost to ...
0
votes
0answers
13 views
To what extend can pre-training and training affect the results of a prediction model?
If we pre-train a model forecasting COVID-19 with data of SARS, which had a different transmission pattern, will our model be weakened?
If we train the same model with data of (for instance) the USA, ...
0
votes
1answer
55 views
Training of a deep Artificial Neural Network
I have few doubts related to training a neural network with more parameters (weights and biases) than number of data points. I know there exists discussion (on this platform) related to training such ...
0
votes
1answer
43 views
Why does shuffling in train test split have a big impact with my loss and accuracy?
I used Keras for the train test split.
This is what I get when I shuffle during my train test split:
When I disable shuffle by setting shuffle:False this is what I ...
0
votes
0answers
24 views
svm train score is 1.0 and test score 0.996. How to verify its correct?
My dataset has features CO2, PM2.5, Temperature and Humidity and I am trying to predict the Air Quality Index based on this information. I have approximately 24,000 data points for each of the ...
2
votes
1answer
28 views
Is it possible to overfitting within single epoch
Let me put my question first. For a time-series prediciton, is it possible to overfit even within the first epoch, when training data and validation data should all "new" to model?
Features and ...
0
votes
0answers
45 views
Machine Learning: Why do I have this pattern of train and validation accuracy?
I am trying to understand what would generate this pattern of accuracy in train and validation dataset (second and third plot below).
I am training a network to recognize 6 types of faces (they are ...
0
votes
0answers
55 views
Can the size of the training set of a neural network be smaller compared to other models?
If some of you knows what is the rule thumb one in ten rule this what I want to discuss.
Neural network can, nowadays, go deeper in performances. Did you find, according to your experiences, that you ...
1
vote
2answers
17 views
Issues with training on a sample of training set?
I am training an SVM on highly imbalanced data. I have rectified this issue and my ML pipeline works just fine. I have allocated 70% of my dataset for training, however this takes an infeasible amount ...
0
votes
0answers
15 views
How to improve neural network training against a large data set of points with varying magnitude
I am currently using TensorFlow and have simply been trying to train a neural network directly against a large continuous data set, e.g. $y = [0.014, 1.545, 10.232, 0.948, ...]$. The loss function in ...
3
votes
1answer
167 views
What is the relationship between mean squared error and classification error?
I've trained a network using a genetic algorithm and I have two possible fitness functions for my GA: MSE and CErr.
If I use MSE as my fitness function, over time MSE decreases and classification ...
3
votes
1answer
614 views
Keras: What is the meaning of batch_size for validation?
If I understand correctly that batch size is the number of samples used in the training of a NN before the gradient gets updated, then why do we need a specified batch_size for the validation sample?
...
0
votes
0answers
14 views
Can I stop training my neural network at this point?
When the validation error of my Neural Network that I am trying to train is slowly decreasing but not by much, is it okay to stop train the network at that point, or do I need to increase the training ...
0
votes
0answers
40 views
How to estimate the performance of Neural network
I have a feedforward neural network with two hidden layers built in keras.
let say I have 40 observations. I split the data into train (e.g., 35 observations) and test (e.g., 5 observations) sets. ...
1
vote
1answer
33 views
What are the best models/methods for training when the target is multidimensional?
I am not very familiar with all methods in Machine Learning. However, I know for example when I apply linear regression, the y is always assumed to be one-dimensional.
My target is multi-dimensional. ...
0
votes
0answers
9 views
Train a neural network as part of other ML algorithm
I want to implement some kind of neural network as "encoder" that encode my input, transform it into other dimension and use the output (encoded data) in other algorithm like kNN etc.
My question is, ...
0
votes
0answers
28 views
Increased computation time for training and prediction with reduced feature space?
I implemented a PCA algorithm to reduce the input feature space of my neural network from 230 to 110 features.
My naive expectation was that if I train a neural network using the same hyper ...
0
votes
0answers
46 views
Data leakage: Does it make sense to split time series this way?
I’m trying to make sense of how to create the training data for time series.
Let’s say we have sales data:
...
1
vote
1answer
31 views
Cross validation in trainControl function
If we use cross validation in trainControl function, still do we need to perform the prediction on test set or training data in train function is sufficient?
I split the data in training and testing,...
1
vote
0answers
42 views
is training dataset of machine learning are Big Data processing results? [closed]
I need to understand the application of machine learning in big data processing. I am so confused with the concept of big data processing and machine learning and I need clear responses. In general, ...
1
vote
0answers
23 views
Model loss stays the same for hours before dropping
I'm training a CNN to colorize images. The model I have is not incredibly deep, and should work fine on the card I'm training on (2080 TI). Initially, I suspected the model was flawed in some way ...
2
votes
0answers
24 views
Goodness of Fit Test vs Testing RMSE
I have a bit of a broad question. It seems to be that there are two different approaches (Borrowing a bit of Breiman's Two Culture's paper) when it comes to testing if our data looks like our model.
...
0
votes
0answers
84 views
Oversampling/Undersampling in respect to Train and Test - Isolation Forest
I've got a quite imbalanced data set.
144.496 : 162 -> ratio of 1000:1
I would like to use IsolationForest to detect the 162 anomalys.
I've already split the data.
However, the iForest doesn't ...
0
votes
1answer
28 views
Validation loss is decreasing, accuracy is decreasing too
So, I have the following charts from my experience.Can any one explain why accuracy is decreasing while the loss in train and validation is decreasing? The point is that i can't early stop too in the ...
0
votes
0answers
59 views
object detection loss
I have trained an ssd detector in my own dataset and the values of train loss and val loss are shown in the picture. However in all the epochs the value of val loss is lower than that of train loss ? ...
2
votes
1answer
25 views
Should a model be trained until it is stable to find optimal hyperparameters?
A model may take several days to train until it reaches an equilibrium - say if the change in error between epochs is lower than some threshold $\epsilon$, or accuracy reaches some equilibrium.
When ...
2
votes
1answer
50 views
should I re-initialize my optimizer and my scheduler before I try to fine tune my neural network on the different dataset?
I am doing NLP, and I have this block of Transformer body that was already trained on dataset A.
Now I am interested in fine tuning this same Transformer on a new dataset B.
In my Python code, should ...
