Methods and principles of selecting a subset of attributes for use in further modelling

learn more… | top users | synonyms (2)

0
votes
0answers
36 views

On population variable importance

Consider we run a random forest on $n$ independent realizations of a random vector $(X_1,X_2,X_3,Y)$ assuming $Y$ is a numerical response variable. Let $f$ be the best theoretical classifier defined ...
0
votes
0answers
27 views

Featuring Engineering from Trends in the Training Set

I have a predictive model with just OK performance, and I'm trying to improve it with feature engineering. My question: is it valid to create new features by looking at trends in the training set? ...
0
votes
0answers
13 views

When to cluster features for supervised learning?

I'm doing a project on dog adoption patterns, and I realized that there are many (100 +) different breeds of dogs. I'd like to build a predictive model using breed as covariate, but I'm not sure ...
2
votes
2answers
89 views

What interactions to include in my GLM model?

I realize this might be a too general question, so I'll describe what I'm doing right now first. I'm working for a virtual insurance company and I have this dataset. It has severity (meaning ...
0
votes
1answer
190 views

Caret: customizing feature selection, nested inside cross validation

Using caret, I want to train a SVM classifier and estimate its performance using repeated cross validation. My dataset has a very large number of predictors (300K) and I want to reduce this number ...
0
votes
0answers
6 views

Design a feature with time and presence information

Context: I am working on a decision tree classifier, trying to classify businesses as to whether they are likely to have an event occur (default) in the next 90 days. One input I get is whether, and ...
0
votes
2answers
305 views

How to normalize time series?

This is a general question on normalization of data so that all the variables are within the same range. Why do we normalize data in pattern classification? How to normalize time series which is ...
0
votes
2answers
53 views

Feature Selection using (low) MCC

I have approximately 1200 input parameters that I am trying to whittle down with the following rough process: 1) Fit rbf SVM with n = 1200 parameters and calculate Matthews Correlation ...
2
votes
1answer
26 views

Boruta score goes to minus infinity

I'm running the Boruta algorithm with a $179\times 36$ predictor matrix and a numerical response. Most of the variables have a score going to -Inf. Should I ...
1
vote
0answers
18 views

How to assess the importance of the features which come from intersection of features of the two models?

I have two models from two different data sets. Model 1 contain 50 features and model 2 contain 40 features. the intersection of features of model 1 and 2 is 10. so how can I assess the relative ...
0
votes
0answers
32 views

How many features can be used for classification?

Asked a similar question the other day without an answer Link. I think maybe the question there is too big. Here I want to ask a specific one: 2 Class labels (Binary classification labelled with ...
0
votes
1answer
26 views

Including or discarding a variable based on its frequency

I am trying to dummy code a variable sector which have 20 levels. I am trying to combine some of levels. However I got this situation, if I have a say 2000 observations, how do I decide on the minimum ...
2
votes
3answers
549 views

Methods in R or Python to perform feature selection in unsupervised learning

What are the available methods/implementation in R/Python to discard/select unimportant/important features in data? My data does not have labels (unsupervised). The data has ~100 features with mixed ...
1
vote
1answer
76 views

Feature Selection - Mutual Information with response variable that takes three values

I am trying to calculate Mutual Information scores for Feature Selection. I have successfully implemented the Mutual Information to test each feature against the binary response variable. Each ...
0
votes
0answers
50 views

How many features to overfit the classification?

Recently, I have got some 'strange' comments from the reviewer of my paper. In my paper, I discussed a novel feature extraction method, and then I compared three classification methods for my binary ...
1
vote
1answer
258 views

Why does SAS Enterprise Miner keep all dummy variables for a coded categorical variable in stepwise logistic regression?

SAS Enterprise Miner nicely creates coded dummy variables for any categorical variables used in a logistic regression model. When it performs a variable selection using stepwise sequential selection ...
1
vote
0answers
51 views

Using Leave-One-Out Cross Validation with LARS

I have a kind of obscure question about using the Least Angle Regression (LARS) algorithm for variable selection. If I'm understanding it right, my professor formulates LARS as such: $$\mathbb{min}\ ...
1
vote
0answers
21 views

Setting up feature vectors

I am working on a classification project and I want to use SVM's and/or Clustering Algs. What I am having trouble with is deciding how to set up my feature vectors. I have already decided what my ...
0
votes
0answers
13 views

Feature selection: all features vs a subset of them

I am doing a binary classification. The dataset has 3000 samples, and each sample has 10 features. But I find that the performance of using all 10 features is almost the same as that of using only the ...
0
votes
0answers
111 views

How to interpret merits in Weka with ChiSquaredAttributeEval and SVMAttributeEval?

I want to interpret the goodness of attributes using feature selection with 10-fold cross validation. With ChiSquared I get something like this (deletet attributes with merrit was 0 in all folds): ...
2
votes
0answers
53 views

category selection with LASSO

Suppose one has two features: color = {R, G, B} and t-shirt size = {S, M, L} and wants to regress these features on the probability of a sale, call it p. So the model is p ~ color + size. Now, the ...
0
votes
0answers
39 views

Feature selection from wavelet transformation in R

I am new to wavelets. Currently, I am developing a prediction model using time series data. I am using the wavelets package in R. I am taking part of the time ...
1
vote
1answer
131 views

Maximum Entropy Model for classification, what to use as context & feature?

I'm building a Maximum Entropy Model to classify some text, based on paper "A Maximum Entropy Approach to Natural Language Processing" by Berger et.al. It's similar to POS tagging. Below is some ...
0
votes
1answer
114 views

Which feature selection method to use for classification problem

I have to do some feature selection for a classification problem with numeric features. I am not sure which feature selection method to use. Chisquared test or Spearmann's rank correlation ...
1
vote
1answer
1k views

Feature selection : how to select the Information Gain threshold?

I am trying to use Information Gain to select features when classifying text with a Support Vector Machine. For each word in our training data, we computed its information gain. Then, we should keep ...
0
votes
1answer
115 views

Feature Normalization/Standardization before or after Feature Selection?

Should the process of feature normalization/standardization be done before or after the feature selection process?
0
votes
0answers
186 views

recursive feature elemination in R with caret

i work with R caret software package to select the most important features from some set of data. My response is a factor of multiple classes (e.g. nominal ...
0
votes
0answers
16 views

Nonlinear functions of other features as new features in SVM model with RBF kernel

Can some one give me some conceptual insight on the potential advantages of disadvantages of adding features that are (nonlinear) functions of existing features in training an SVM model with an RBF ...
2
votes
2answers
223 views

Model Tuning and Model Evaluation in Machine Learning

Despite my readings (on stack 1, 2, or in literature (Cawley, 2010; Japkowicz, 2011)), I don't find a clear procedure for tuning and evaluating a model in a classification task. I want to perform a ...
1
vote
0answers
77 views

Unstable models, repeated crossvalidation, feature selection

I'm still trying to classify few (about 200) samples in a high dimensional feature space (dim=19) into 3 (very unbalanced) classes. I use an implementation of Least Squares SVM with one vs one coding ...
1
vote
2answers
772 views

Best feature selection method for naive Bayes classification

i want to make classification with naive Bayes. I have got about 100 Features. Numerical ones as well as categorical ones. Since i want only the most relevant ones to be included for the ...
0
votes
0answers
75 views

TF-IDF for text classification by taking into account the document class

I am looking for a TF-IDF weighting for text classification (not document ranking/retrieval) which takes also into account the document class. For example let's use the typical spam/not spam ...
1
vote
0answers
29 views

Removing insignificant variables? [duplicate]

Suppose you fit a linear regression model on some data with 10 variables. The F-statistic shows that 3 of them are significant (p < 0.05) , 2 are within trend (0.05 < p < 0.10), and the other ...
1
vote
0answers
88 views

Is this normal to have big value of Chi Squared?

I am using chi squared for feature selection in text classification. However when I compute it I sometime have very big values. Like 100, 1000 or even 20000. Is this normal ? I wonder because I ...
2
votes
1answer
327 views

Random forest cross validation for feature selection, imbalanced datasets

I have an 5297X26 imbalanced dataset, the class1 has 588 samples and class2 has 4709 samples. I used the following code to perform random forest: ...
1
vote
0answers
145 views

How to apply feature selection based on tf-idf threshold

Let's say we have the following matrix (typical VSM example): ...
0
votes
1answer
183 views

Find variables selected for each subset using caret feature selection

I am doing feature selection using the command 'rfe' in the caret package (http://caret.r-forge.r-project.org/featureselection.html). This command uses a metric to find the optimal amount of variables ...
0
votes
0answers
92 views

Dropping predictor variables, based on variable of importance, effect of Random Forest Accuracy

I am trying to use Random Forest to accurately predict forested land cover classes using Landsat 7, climatic and geographical data. I have 23 predictor variables and 1 response variable. When I drop ...
3
votes
2answers
248 views

Feature / attribute selection for k-means or other clustering

It seems to me that in literature it is assumed that one knows which features / attributes to choose to characterize an item in clustering. If I have a database with items which have many attributes, ...
0
votes
1answer
30 views

Importance of Time Features

if you have a time series and you want to do some predictions, what time feature should you use ? lets say we are trying to predict how many people visit a certain website, we have data for the ...
4
votes
2answers
177 views

Feature selection with partial permutation

For feature selection, permutation tests are biased in favor of those categorical variables with a large number of levels [White1994]. Besides, it has been proposed [Deng2011] that partial ...
0
votes
0answers
54 views

Regression by multiple dependent variables with constraints & feature selection

I have a data set of 1000 records. Each record has three dependent variables $y_1, y_2, y_3$ and 100 independent variables $x_1,...,x_{100}$, where the dependent variable $y_i$ satisfies: $0\le y_i ...
2
votes
2answers
503 views

How to select a subset of variables from my original long list in order to perform logistic regression analysis?

My situation: small sample size: 116 binary outcome variable long list of explanatory variables: 44 explanatory variables did not come from the top of my head; their choice was based on the ...
2
votes
1answer
59 views

why feature scaling or weighting is important in surpervised learning?

I can understand feature scaling or weighting is important in unsurpervised learning case, because we want an good representation of "similarity". But why it is also important in surpervised learning ...
2
votes
0answers
62 views

Feature selection: permutation test Vs deleting a variable

In feature selection for predictive models, it is usually applied a permutation test. In this test, all the values of one variable are randomly permuted and the prediction accuracy is extracted for ...
0
votes
0answers
26 views

Comparing F-values of covariates in R

I'm new here and have a question regarding ANOVA in R. I have an ANOVA table like this from running anova(model) in R, where ...
1
vote
0answers
39 views

Is linear kernel SVM performance between features indicative of RBF kernel SVM performance?

I have feature set 1 and feature set 2. If a linear kernel SVM performs better ("better" meaning greater classification accuracy) when using feature set 1, does this guarantee that a properly tuned ...
1
vote
1answer
31 views

Transform a non-monotonic value before decision tree (concrete example)?

Newbie question here. I am building a toy decision tree to differentiate personal names from business, government, or organizational names, like: AAA ENTERPRISES LLC DBA AAA BBB SERVICE SMITH ...
0
votes
0answers
47 views

What method should I use for this optimization / feature selection project

I'm going to describe a problem and I'm not sure how to best solve it. I will describe the situation. When answering please recommend a method and maybe a software library. I'm using Python for my ...
0
votes
1answer
109 views

Values of the weights in Adaboost

I have implemented a simple Adaboost algorithm, using several weak classifiers, and when checking the values computed by it there are alphas with a negative value. Is that possible, or is there a bug ...