soufanom
  • Member for 9 years, 3 months
  • Last seen more than a month ago
When does Naive Bayes perform better than SVM?
27 votes

There is no single answer about which is the best classification method for a given dataset. Different kinds of classifiers should be always considered for a comparative study over a given dataset. ...

View answer
Feature selection with Random Forests
13 votes

For feature selection, we need a scoring function as well as a search method to optimize the scoring function. You may use RF as a feature ranking method if you define some relevant importance score. ...

View answer
Handling unbalanced data using SMOTE - no big difference?
9 votes

I would like to bring to your attention also that in the original SMOTE paper, the good results were based on both combining SMOTE and random under-sampling. This is because applying SMOTE to achieve ...

View answer
Increasing the sample size does not help the classification performance
Accepted answer
5 votes

Increasing training size does not neccessarily help the classifier and rather, may lead to a degradation in the generalization ability. Regarding your own experiment, the factor of such unexpected ...

View answer
Test for linear separability
Accepted answer
4 votes

Well, support vector machines (SVM) are probably, what you are looking for. For example, SVM with a linear RBF kernel, maps feature to a higher dimenional space and tries to separet the classes by a ...

View answer
Predicting chemical property (Boiling Point) from a SMILES string
Accepted answer
3 votes

It is a matter of generating features or variables that describe the SMILE representation of a chemical compound. Computational chemistry has proposed good definitions of different chemical ...

View answer
Defintion for model diversity?
Accepted answer
3 votes

There is a great book on the topic of ensemble classifier. It is online on: Combining Pattern Classifiers There is a full chapter (ch10) in this book on diversity and how to measure it. A set of ...

View answer
Exploratory analysis: regression model with mutually correlated predictors to explain a dichotomous outcome?
2 votes

Your approach seems valid for me. However, it does not consider conditional or interdependent relationship of variables within every group. Your are building a regression model for every individual ...

View answer
Is this training dataset enough for training and testing classification model?
2 votes

From my experience, I would answer "Yes" to your question although a wise one would be "It depends". You may refer to the following other threads for extended relevant discussions: How large a ...

View answer
The usage of data mining in pharmaceutical companies?
2 votes

As you may know, "Data Mining" is a term that can be viewed as an overlap between "Databases" and "Machine Learning". Both are being exploited into the drug discovery field. Also, traditional mining ...

View answer
How do important and insignificant variables impact model?
2 votes

I would like to go with the advices @MattKrause with more highlight on the feature selection part. Once you check the business rules and generate all the representative variables, you are advised to ...

View answer
Is it needed to normalize data before rule model extraction algorithms like ID3?
Accepted answer
2 votes

In general, if you have nominal valued features, then there is no need to normalize data for both Naive Bayes Classifier (NBC) and Decision Trees. This is because we are dealing with discrete ...

View answer
What kind of general strategy can you apply after selecting model and hyper parameter training?
Accepted answer
1 votes

This is a very general question about enhancing performance of a machine learning model. It is always data dependent to decide on the best approach to improve precision or performance. As you point ...

View answer
Can someone explain to me the Bayesian classification model?
1 votes

The Bayesian approach proposes using our previous expereince or knoweldge, if available. For example, If I ask you what is the probability of having a head in a coin flip, you will probably say 50%. ...

View answer
Multivariate Bayesian formula
Accepted answer
1 votes

Intuitively, you may see the result of this formula by drawing the corresponding graph as follows: The joint distribution represented by this graph is given by p(c|a)p(b|c)p(a). Now, summing out c, ...

View answer
What are the general strategies in creating a Probabilistic Graphical Model?
1 votes

I will try to answer some of your questions here regarding Probabilistic Graphical Models (PGMs). Before I start, an excellent course to follow is the one offered here by Daphne Koller and her book ...

View answer
How to statistically compare two algorithms across three datasets in feature selection and classification?
1 votes

You are running featuer selection with GA 10 times and every time you get a different output !! First, If you start by the same seed, you should always get the same selected featuer subset. However, ...

View answer
Classifier feature importance
1 votes

The wrapper model popularized by Kohavi as mentioned @Peter would help in finding optimal features which are not necessarily relevant to your target labels. In the same paper, Kohavi states that "...

View answer
When is the PMI value good or bad?
Accepted answer
1 votes

As mentioned in Wikipedia, "PMI is zero if X and Y are independent and PMI maximizes when X and Y are perfectly associated". It is also mentioned that "Pointwise mutual information can be normalized ...

View answer
Changing variable values and examine the outcome difference between the altered and original data
1 votes

I would judge this method as a sound one if we know their aim of doing such manipulation to the variable. However, I can share two cases in which such a procedure is sound. Case 1: Improving ...

View answer
Fisher Distance for feature selection
1 votes

You may experiment with two representations of the problem. 1- Group ECG channels of one class together and label class A as 1 and class B as 2. So, the target variable is composed of two classes ...

View answer
An intuitive meaning of the area under the PR curve?
0 votes

Well, I will try to give some intuition close to the one of Wikipedia as you may desire. The PR-AUC can be thought of as the probability that a classifier will rank a randomly chosen "positive" ...

View answer
Determining conserved features using a Bayesian approach
0 votes

Basically, two approaches are mentioned in your post: 1- Using single FS metric over the entire training dataset 2- Partitioning data and using the FS metric over every split In the second case, ...

View answer
Can feature selection be considered a way to observe relationship between variables like correlation?
0 votes

A variable that achieves a better classification performance is not necessarily a variable that is correlated or relevant to the target class labels (i.e. Y in your case). So, optimality does not ...

View answer