Variables (used for prediction or explication) used in regression or regression-like models (like clustering, discrimination). Use this tag for questions about constructing such variables or selecting the best among them.

learn more… | top users | synonyms

0
votes
1answer
12 views

DateTime Variable — Kmeans

Suppose I have a DateTime variable that I want to cluster on. I only really care about the day, hour, and minute. Is it wrong to create a column in my data set for each of these or should I keep the ...
2
votes
1answer
46 views

Can I use output of classifier A as feature for classifier B?

This is likely to be a confused question, but I'm curious if this is a valid way to combine classifiers. I have a classification data set, i.e. column of labels and N columns of features, and I use a ...
1
vote
1answer
28 views

FInding relevant features for a time series segmentation

I have a time series data, where each of the data point belongs to one of the known clusters. What I am interested is to perform a HMM so that we can obtain hidden states that further abstracts out ...
0
votes
0answers
39 views

time stamp as input variable for regression (feature extraction)

I am working on web logs and have a time-stamp variable in the format dd-mm-yyyy hh-mm-ss. I have earlier worked on date variable and found that best way to extract feature from date is to create ...
0
votes
1answer
25 views

Letter recognition feature extraction

I have few thousands letter's images which are all colored black on a white background. I wanted to extract their features so that it was possible to cluster them effectively. My attempt was to ...
1
vote
0answers
7 views

How flexible are autoencoders for non-linear dimensionality reduction?

I've been starting to play around with autoencoders for feature extraction and dimensionality reduction, and am wondering how critical input feature definitions are for success. For example, some of ...
-3
votes
1answer
44 views

Why should I choose features or plot them manually when there are built-in functions to do that?

Why should I select variables due to my intuition if there are builtin functions in sklearn python like SelectKBest() and PCA() If I plot graphs of features of the data to see if they can detect the ...
1
vote
1answer
23 views

How to deal with a variable-sized real vector of inputs?

I have a collection of objects with properties that I measure. For each object, I obtain a vector of real numbers describing that object. Each object results in a vector having a different length. I ...
0
votes
1answer
32 views

How to get the best out of a “bad” set of features for regression?

I'm trying to learn a regression model for a computer vision / pattern recognition task, where I try to estimate a continuous variable from a set of visual features. I have done preliminary ...
0
votes
2answers
36 views

How to normalize feature vectors for concatenating

I have two different feature vectors of completely different scale, which are to be used as training data for machine learning algorithm. When I concatenate them, should I scale and normalize them ...
0
votes
1answer
28 views

General Criteria for Feature Quality

It is generally accepted that the most important factor for successful machine learning is quality feature engineering: Feature Engineering is the Key At the end of the day, some machine ...
0
votes
1answer
16 views

How to deal with quasi-continuous features?

I've searched around a bit for strategies to approach the problem I'm facing and haven't come up with much. I'm working with a data set that has many "quasi-continuous" features. That is, the ...
0
votes
0answers
15 views

Should one drop features not returned by xgb.importance?

The xgb.importance() function within R package xgboost was used. Have some questions in mind, and cannot seem to find a direct answer somewhere else, if someone can address these: When xgb....
0
votes
2answers
48 views

Is feature engineering relevant at all for Random Forests?

Random forests is an ensemble of trees that learns the hidden patterns in the data. I have mostly tried doing some feature-engineering before running the Random Forest model but is it required or the ...
0
votes
0answers
27 views

Part-of-speech tags as Document Term Matrix

For my thesis I need to apply a part-of-speech tagging for sentiment classification in R. I have a dataset consisting of ~800 sentences which were tagged by the ...
0
votes
0answers
16 views

Feature Matrix vs Feature Vector

In my application I have essentially $n$ areas of interest in an image. The image is circular in nature so the zones are slices. Each zone has 7 features associated to it. The goal is to detect ...
0
votes
0answers
9 views

How to generate count-based features from categorical data for binary classification?

I recently discovered this blog post by Microsoft Azure. In it they describe a method of generating new count-based features from categorical features for a binary classification task. I am a bit ...
0
votes
0answers
17 views

Rescaling vs Standardization of features

Is there any general rule of thumb or any justified rule to choose whether to scale a dataset using Rescaling (for each feature, subtract the min value and divid by the max - min) or Standardization (...
1
vote
2answers
43 views

Can trees or random forests learn ratios

This is a question about feature engineering for decision trees/random forests. Given two continuous variables X1 and X2, is it ...
1
vote
0answers
28 views

Encode a tree into a machine learning feature

I am creating/working with a dataset in order to answer all kind of questions using machine learning algorithms. One specific issue is that I would like to create a new feature based on a tree ...
4
votes
1answer
49 views

Do word vectors obtained via word embedding techniques really form a vector space?

Word embedding refers to feature learning techniques in natural language processing where words are mapped to vectors of real numbers in a low-dimensional space, the embedding space. Similar to ...
0
votes
0answers
40 views

How to transform feature with peak at zero to normal distribution?

I have a feature in my dataset which has lot of zero values, i.e. a big peak at zero (the zeroes are valid and valuable information). The histogram is the following: I want to transform all my ...
0
votes
0answers
9 views

Difference between polynomial regression and polynomial kernel

A few answers on SO suggested that a polynomial transformation and a regularized regression can be used instead of a polynomial kernel regression. What's the difference between them? I thought ...
0
votes
0answers
18 views

Create features from a document

I have been given an assignment related to NLP and I am a newbie in this field. Train a named entity recognition system that treats the documents as strings of mentions, x . A labelling of the ...
4
votes
3answers
384 views

random forest - summarize two features in one without losing information

I am training a random forest on a dataset including both categorical and numerical features. In particular I have a binary feature, call it $x_1$, which has $0$ or $1$ as possible outcomes. I also ...
3
votes
0answers
16 views

Extract a sub-distribution with specific characteristics

Imagine that I have a distribution of some data. For example a distribution of natural numbers [1, 10, 45, 89, 12, 9, 4, 100]. I characterize such distribution with 2 features: the mean = 33.75 and ...
1
vote
1answer
28 views

How to upweight an arbitray feature in GBM?

Are there any ways to upweight any particular feature for GBM (tree boosting)? The motivation is: Assuming one would like to put a few dummy indicators, e.g., X_A, X_B, X_C, ..., into GBM, for the ...
0
votes
0answers
17 views

Input matrix construction for time series

I am having problems constructing the input matrix for a data analysis / machine learning task. The data set consists of ~300 data points, each one in form of a matrix where rows are time steps, ...
4
votes
1answer
134 views

How to choose/train Matrix Values of Convolution Kernel in Neural Networks

I'm trying to better understand convolutional neural networks better by writing up Python code that doesn't depend on libraries like Convnet or TensorFlow, and I'm getting stuck in the literature on ...
0
votes
0answers
25 views

Multiclass classification with large number of classes but for each user the set of target classes is known

I'm new here but I'm familiar with some machine learning theory (took some courses in school) and my question is more about how to apply ML in a practical setting. I have this project where I'm ...
4
votes
1answer
127 views

Does the first principal component differ from simply computing the mean of all variables?

I was just wondering if the first principal component, while I am trying to find it for a dataset of 18 variables, is different from simply adding all variables and finding the mean? I.e. to compute ...
0
votes
0answers
195 views

Feature extraction for time series classification

I have electric signals represented as 2(internal)d data. Each data point is in the form (<timestamp>, <power_value>). I want to be able to ...
0
votes
0answers
33 views

Expand or compact features?

I have a classification task for people with 3 categories. I want to apply machine learning for that. I have 10 sources of data, which have the same fields (say 4: age, job title, number of ...
-2
votes
1answer
92 views

Standard deviation vs. variance as a feature

I'm using the median of response times (of users) as a feature in a machine learning context. As a second feature I want to use the standard deviation or variance of the response times. Of course, the ...
1
vote
0answers
39 views

Feature binarization for RF/GBMs?

Are there any advantages to feature binarization for random forests or gradient-boosted machines? For example, suppose I am predicting snowstorms for the next day using various past measurements - ...
0
votes
0answers
14 views

Handcrafted a feature or Convolution nets

I am doing an image classification projects. Image in each class represents the same object deformed with an unknown complicated function. In fact, images in the same class become very dissimilar due ...
0
votes
1answer
34 views

How many features should I say I have in my model?

I am running machine learning using name features to predict Y (binary 0 and 1 labels). Using the name entity (eg: John Carter), I derive into 4 name substring features (1: first name = "John", 2: ...
1
vote
1answer
44 views

How to take the index of the nearest centroid as a feature?

To create additional features for a dataset I have conducted a cluster analysis and assigned a feature to a data set for cluster membership: ...
1
vote
0answers
50 views

Combining categorical and continuous features in DNN

I am creating an application that can take as inputs, two numbers (1 or 0) as well as a class defining a binary operation (AND OR XOR etc) and training the network to preform the operation. Without ...
2
votes
0answers
44 views

How feature stacking works?

I don't understand how features stacking works. I found out the following sample guideline: Split the train set in 2 parts: train_a and ...
0
votes
0answers
14 views

Features work for pairwise binary classification, but fail in three classes

Let $a,b,c$ be three classes. Let $X$ be a matrix of row/column samples and let $F_i$ be several sets of features (i.e. from a sample matrix $X$ we can extract a set of features $F_1$ for each sample ...
0
votes
1answer
27 views

feature hashing as a form of encryption

Is feature hashing a really bad way to do encryption? I want to encrypt my documents but in a format that some ML algorithms can be trained on them one day.
0
votes
0answers
10 views

Missing data and feature scaling

I am interested in doing some feature scaling to try and tease out something from my data (box plots by outcome show that the 25/50/75 quantiles are very similar; certain variables have more "outliers"...
0
votes
0answers
16 views

Correct evaluation/ comparison between undercomplete and overcomplete representations

Suppose I'm performing Unsupervised Feature Learning method to learn a representation of the data that is under-complete (e.g. 100 features) and use another algorithm to learn an over-complete ...
0
votes
0answers
35 views

Automated Feature Construction / Feature Combination in Matlab or Weka

I would like to extract more information and to catch important relationships in my numeric, sequential time series data. I suspect I have not yet sufficiently discovered nonlinear relationships ...
0
votes
0answers
26 views

How to model Not Applicable (N/A) input in a Neural Network?

Assume I have a Neural Network with an input value to distinguish 2 discrete categories, e.g. Bicycle (input -0.9 in case of tanh) or Car (input 0.9 in case of <...
0
votes
0answers
24 views

How do I have select features which are influential for prediction?

I have a dataset which has dependent variable(label) as possible destinations and independent variable(features) as age,language, gender and many other categorical variables. How do i find which are ...
0
votes
0answers
29 views

How to implement Vectors in Feature Vector?

Please change the question title if you find a better one for my problem. I have taken some measurements with a 6 DOF (accelerometer and gyro) sensor. Now I want to use this data for a classifier ...
1
vote
0answers
48 views
0
votes
0answers
50 views

SVM accuracy after log transformation

I am using SVM (RBF kernel, the LibSVM implementation) to deal with a classification problem. When I use a log 10 transformation for may features values instead of using the default scaling method ...