prashanth
  • Member for 6 years, 5 months
  • Last seen more than a month ago
  • India
How to choose the number of hidden layers and nodes in a feedforward neural network?
18 votes

• Number of hidden nodes: There is no magic formula for selecting the optimum number of hidden neurons. However, some thumb rules are available for calculating the number of hidden neurons. A rough ...

View answer
Minimum number of layers in a deep neural network
12 votes

As per the literature, Schmidhuber, J. (2015). "Deep Learning in Neural Networks: An Overview". Neural Networks. 61: 85–117. arXiv:1404.7828free to read. doi:10.1016/j.neunet.2014.09.003. https://en....

View answer
What are real life examples of "non-parametric statistical models"?
8 votes

As Johnnyboycurtis has answerd, non-parametric methods are those if it makes no assumption on the population distribution or sample size to generate a model. A k-NN model is an example of a non-...

View answer
hyperparameter tuning in neural networks
Accepted answer
7 votes

The link provided in @itdxer's comment is great. Based on this link, I am writing this answer. Hyperparameter optimization is neural networks is a tedious job as it contains many set of parameters. ...

View answer
Should we normalize before using VarianceThreshold in sklearn?
6 votes

Yes, one must do normalization before using VarianceThreshold. This is necessary to bring all the features to same scale. Other wise the variance estimates can be misleading between higher value ...

View answer
Why is t-SNE not used as a dimensionality reduction technique for clustering or classification?
5 votes

t-SNE does not preserve distances, but it basically estimates probability distributions. In theory, the t-SNE algorithms maps the input to a map space of 2 or 3 dimensions. The input space is assumed ...

View answer
Generate synthetic data to match sample data
5 votes

I am trying to answer my own question after doing few initial experiments. I tried the SMOTE technique to generate new synthetic samples. And the results are encouraging. It generates synthetic data ...

View answer
Data normalization and standardization in neural networks
5 votes

You could do min-max normalization (Normalize inputs/targets to fall in the range [−1,1]), or mean-standard deviation normalization (Normalize inputs/targets to have zero mean and unity variance/...

View answer
Feature importances in random forest
Accepted answer
2 votes

In Random forest, generally the feature importance is computed based on out-of-bag (OOB) error. To compute the feature importance, the random forest model is created and then the OOB error is ...

View answer
What is the best source to learn Random-forest algorithm in Matlab from scratch?
2 votes

Matlab already provides a function for 'Random forest'. It is the 'TreeBagger' function. More info in https://in.mathworks.com/help/stats/treebagger.html To understand basics of Random forest ...

View answer
Can we predict the categorical variable of the given dataset?
2 votes

Yes, a predictive model can be developed to predict the Survived feature. A problem similar to this is given in the link https://www.kaggle.com/c/titanic. It has some very good tutorials too. This ...

View answer
convolutional autoencoder on an odd size image
1 votes

I got the solution by adding a cropping layer in the end. The full code is as below: from keras.layers import Input, Dense, Conv2D, MaxPooling2D, UpSampling2D from keras.models import Model # from ...

View answer
How do I select right features
1 votes

Feature selection is a more relevant process when there are large number of features (in the order of 1000's) with correlations betweent them and some being irrelevant as well. Here in this case, as ...

View answer
Clustering of mixed type data with R
1 votes

k-prototypes clustering might be better suited here. It combines k-modes and k-means and is able to cluster mixed numerical / categorical data. For R, use the Package 'clustMixType'. https://cran.r-...

View answer
How to compare the F-measure values?
1 votes

To say that M1 significantly improved the result in compare with other methods, you must have to use statistical methods and tell that the measure or difference is statistically significant. To do ...

View answer
Random Forest regression and MSE
1 votes

To check the performance of a split, as you mentioned MSE and RMSE are the popular approaches. RMSLE penalizes an under-predicted estimate greater than an over-predicted estimate ϵ is the RMSLE ...

View answer
Do extra hidden layers prevent convergence?
1 votes

For a deep neural network that you mention, finding an effective local minima is the key. As per the paper, Gülçehre, Çağlar, and Yoshua Bengio. "Knowledge matters: Importance of prior information ...

View answer
Training set, test set and validation set
1 votes

So my question is, why splitting my data at the beginning if I am assessing the performance of each of the 5 models with cross-validation. Every classifier has some parameter(s) to tune. You need to ...

View answer
Boosted Trees classification
1 votes

Perhaps your model is over-fitting. Before you apply the classification algorithm, you need to divide the dataset to training and testing sets. The performance measures for the training and testing ...

View answer
Model Underperforming
1 votes

Your dataset suffers from the imbalance problem. You can approach the problem possibly in two ways: Use sampling techniques to over-sample the minority class (the churn class), or undersample the ...

View answer
Kernel SVM: I want an intuitive understanding of mapping to a higher-dimensional feature space, and how this makes linear separation possible
1 votes

Transform predictors (input data) to a high-dimensional feature space. It is sufficient to just specify the kernel for this step and the data is never explicitly transformed to the feature space. This ...

View answer
How to deal with problems like this ? What machine learning algorithms should be used?
0 votes

Random forests for Regression is a suitable algorithm when there are mixed features (both numeric and categorical). I am assuming you are using Python for coding. Python has sklearn library which ...

View answer
Feature importance in random forest
Accepted answer
0 votes

The feature importance in Random Forest is based on out-of-bag error, and its value is>=0. As per the Wikipedia page, "To measure the importance of the j-th feature after training, the values ...

View answer
Is there a best way to compare different classifiers?
Accepted answer
0 votes

In machine learning, there is no one algorithm that’s always better than others which is as per the “No free lunch theorem”. Therefore, one has to try with different set of classifiers and choose the ...

View answer
Optimal number of clusters using K-Prototypes method in R
0 votes

You may use the code as below to plot the elbow curve. The input to the code below is the . data <- <input the data here> # Elbow Method for finding the optimal number of clusters set.seed(...

View answer
How to choose the best algorithm for measuring attribute importance/relevance?
0 votes

It is generally recommended to use more than one method for feature selection. For instance, taking a combination of outputs from Recursive feature elimination and Random forests can be effective to ...

View answer
Using different classifiers for feature selection and prediction
Accepted answer
0 votes

It is alright to use different classifiers for feature selection and prediction. But generally it is recommended to see the features selected from multiple methods before finalizing the features. A ...

View answer
Using random forest/boosting for classification when important features are known
0 votes

Feature selection is a recommended approach before predictive modeling because the presence of correlated or irrelevant features may make the built model bad. But domain knowledge has its own ...

View answer
What statistical test to use for this scenario?
Accepted answer
0 votes

Repeated measures ANOVA is to be used when the same subjects are used for each treatment. On the other hand, One-way ANOVA is used to test for differences among two or more independent groups (means). ...

View answer
Which one would perform better: logistic regression or discriminant analysis?
0 votes

Do a subject-wise cross validation using both discriminant analysis and logistic regression. Whichever is giving better performance (accuracy, sensitivity, specificity and AUC) in cross-validation, ...

View answer