Questions tagged [feature-engineering]

Feature engineering is the process of using domain knowledge of the data to create features for machine learning models. This tag is meant for both theoretical and practical questions regarding feature engineering, excluding questions asking for code, that would be off-topic on CrossValidated.

Filter by
Sorted by
Tagged with
0
votes
0answers
12 views

Depending samples in ad ranking and click rate prediction

I am struggling with the following problem: Suppose we fit a machine learning model to model advertisers click rates. I used a Logistic Regression approach using a one-hot/dummy encoding. We have two ...
1
vote
0answers
12 views

Feature engineering procedure using optimism corrected bootstrap

I have a dataset with ~600 datapoints, 49 categorical features (five possible categories), and a binary outcome variable. The dataset is incredibly imbalanced, with just over 3% of the outcomes in the ...
0
votes
0answers
14 views

How to handle associated features in machine learning

I am working on a classification project in which some features are linked and I'm not sure how to handle them. I will simplify my project like that : There are different jobs, and multiple ...
2
votes
1answer
26 views

How to use fresh data when target prediction period is long?

I'm using supervised learning on monthly activity data to predict when a customer buys a particular product. This product is typically bought infrequently and at the moment my target variable is ...
1
vote
1answer
35 views

Why Does a Monotonic Transformation Of Dependent Variable Change Variance Explained In Random Forest

I am working with the Boston data set in R. I have read that random forest should be able to deal with untransformed data. In my example I do a log transformation of the dependent variable. My ...
1
vote
0answers
18 views

Polynomial Regression and Feature Transformation

According to Polynomial Regression concept, high order terms in a model such as $x_1^2, \,\, x_2^3, \,\, x_1^2x_2$ are replaced with new features. In this way, the model equation is converted to ...
1
vote
2answers
43 views

How to handle missing data in timeseries classification?

I'm using 16-channel, 400-Hz, standardized EEG data to train CNN-LSTM for seizure classification. The data contains periods of no recordings (flatlines) - spanning anywhere from seconds to minutes - ...
0
votes
0answers
8 views

Tensile test results

I am currently writing my thesis, where I have to investigate if a surface treatment increases the adhesion (between tape and a plastic surface). For this purpose, I have among other tests, performed ...
0
votes
0answers
11 views

Using groupwise averages of dependent variable as predictors

Let's say I have a binary classification problem with a number of categorical predictors. Is there an issue with creating new features by calculating the group wise average of one or more of the ...
0
votes
0answers
11 views

How important are hyper-parameters in SVM based RFE feature selection?

How important are hyper-parameters in svm based RFE feature selection? For feature selection using RFE (recursive feature elimination / selection), I have seen some publications where only "external" ...
3
votes
1answer
42 views

Encoding IP Address as a Predictor in Machine Learning

Is there some approach to "encoding" IP Address (IPv4) in a way that the new representation can capture both cardinality and the statistical distribution of the full range of IP address and also ...
0
votes
1answer
13 views

Feature selection on full training set, does information leak if using Filter Based Feature Selection or Linear discriminate analysis?

In order to test a potential classification set, usually some data is kept as a holdout set, and not used for inner-cross-validation or model training. However, what happens if too many features ...
0
votes
0answers
17 views

Subtraction between features before performing PCA and training a classification model

In my dataset, I've created moving averages for historical indicators. Is it pertinent to perform subtraction between those moving averages before PCA? My stats/calculus feeling: as PCA is a ...
0
votes
0answers
16 views

Model Based Feature Selection vs Wrapped Method Feature Selection

I read about Wrapped Method Feature Selection, I get that it is to look at the features then test them against the predictive model that we need then find if it has an effect or not and then decide to ...
0
votes
0answers
9 views

Feature engineering on hierarchically structured data

I have a regression problem to predict $y$ and a predictor categorical variable $x$ that has about $500$ different categories. $x$ is an occupation code of a company and is hierarchically structured, ...
0
votes
0answers
12 views

What is the methodology behind Filter Based Feature Selection (i.e. Pearson correlation, etc.) on Azure Machine Learning Studio?

Filter Based Feature Selection on Azure Machine Learning Studio supports feature selection and ranking through Pearson Correlation, Kendall Correlation, Spearman Correlation, Mutual Information, Chi ...
1
vote
2answers
91 views

How to calculate Helmert Coding

I am trying to understand how Helmert Coding works I know it compares levels of a variable with the mean of the subsequent levels of the variable, but what are these levels and how can I calculate ...
5
votes
1answer
185 views

How would someone use curves as an input to a supervised learning model?

I was asked this question during a test and couldn't figure out the answer: You have a set of curves against time $X_i(t)$ that you want to use as input to a supervised learning model. The curves ...
0
votes
0answers
13 views

Data Leak during data acquisition for credit scoring

I have a few questions about data leaks. Particularly, I'm interested in a credit scoring data can have leakages. I'm at the stage of data acquisition and I suppose I have target leak but not sure. ...
0
votes
0answers
14 views

Manually creating target variable, getting f1 score 1

I am building a classifier for user engagement in my website. Basically, since there are no "proxy" for engagement, i.e. there is no pre-defined target variable, I came up with minimum thresholds ...
0
votes
1answer
26 views

What is meant by Low-Order combination of features?

I came across a Machine Learning paper that talks about input with low-order combination of features. A statement says: The initial feature is used as the input of the model, and the non-linear ...
0
votes
0answers
15 views

Feature analyzing methods

I am a little confiused how to interpret following situation: I am trying to implement a image classification task using hog+SVM. For that i tried to analyze and understand the properties of the ...
0
votes
0answers
8 views

Avoid learning certain known features / selecting alternative features in neural network training

In an application of neural network to a classification problem, often time one trains the network to pick out different features in the input data set (learnt by the hidden units) and classify the ...
0
votes
0answers
6 views

Speaker normalization of features before model training

I am building a model using a supervised machine learning based on features I extract from speech signals. The features include MFCC, auto correlation and energy derivatives. According to this paper,...
1
vote
1answer
69 views

One hot encoding vs apply the average of the label to each category

I have a fairly reasonably sized dataset (row>50k). And I'm looking for the best way to utilize some of the categorical columns. For purpose of this question, let's say that one of the categorical ...
0
votes
0answers
24 views

How to handle order invariance (and variable length) of certain sample features in machine learning input vector?

Looking into what can be done (or if it is even an issue) when a sample vector xm contains a variable length subvector of features that are similar and order invariant, so sample vector would have the ...
0
votes
0answers
21 views

Performing significance test with respect to cross validation

While performing sentiment analysis, I am trying to assess whether my approach using a novel feature set (similar to the delta-idf technique) outperforms the tf-idf metric using significance analysis. ...
3
votes
1answer
94 views

Is it advisable to use output from a ML model as a feature in another ML model?

Can I use the probability score generated from a Machine Learning model as a feature in another model? For example, say we have a model which generates the probability of an ad being bad. Lets call it ...
4
votes
0answers
32 views

Feature engineering for sheet music

I have a large dataset of digitized music scores that I'd like to use as input to a network. Initially, I'm looking to train networks to identify key signatures, tempo, dynamics, etc. from the raw ...
1
vote
0answers
13 views

Sterling vs Combustion Engines [closed]

This may be difficult to answer but, theoretically, if you had a car with a sterling engine that produces say approximately 300 horsepower how much of any specific heat source would need to be used if ...
1
vote
1answer
33 views

Feature engineering using the target/dependent variable

I am a beginner and my question relates to feature engineering. My task is to help develop a model which predicts whether a customer request is a fraud case. A variable in the dataset is the ...
0
votes
1answer
46 views

Are legacy values useful for regression models?

I'm building a model that predicts house prices in order to learn some regression techniques. Currently I'm trying to engineer features that might be significant when predicting prices. I got a hold ...
0
votes
1answer
39 views

On starting feature engineering

I would like to start my feature engineering process by first selecting a subset of features that are highly correlated with the target feature. However, if I do select let’s say the top k in terms ...
0
votes
2answers
94 views

Feature selection using PCA for linear regression

I am using PCA to the training data set to do feature selection before applying linear regression to build a classifier model. In this scenario, would it be useful to use ridge regression to ensure ...
0
votes
0answers
18 views

Row aggregation of multiple records

We have a syslog dataset with different Timestamps and 3 other features pertaining to syslog information such as Process, Trace information, SeverityType etc. Below is the dataset format with ...
0
votes
0answers
10 views

Feature engineering for verb\non-verb classification

Suppose we have data of X = words, and for each word we have a label indicating whether the word is a verb or a non-verb. So, y = labels. Assuming we can build all the unigrams and bigrams of each of ...
0
votes
0answers
15 views

Predicting Transformation

One of my friends found the paper An Empirical Analysis of Feature Engineering for Predictive Modeling. We were discussing the models the author wrote about. He wrote: If the machine-learning model ...
0
votes
0answers
16 views

Assigning weights to the features used in content based recommendation

I am trying to make a recommendation engine for book business which has following features associated with the books: Book Region Book Market Segment Publish Date Book Genre Book Type and so on ...
3
votes
2answers
64 views

Strange encoding for categorical features

I am reading through https://arxiv.org/pdf/1609.06676.pdf which presents an extension of the isolation forest algorithm so that categorical features may be taken into account. On page 5, the authors ...
0
votes
1answer
18 views

How to deal with varying number of intervals and hence varying number of features dividing an audio signal while classifying these audio signals?

I've $2000$ audio signals, each divided into a number of time intervals/time frames of $50$ miliseconds (ms) and these signals have overlaps for $25$ ms. Now, the audio signals being of different time ...
0
votes
0answers
96 views

Target Encoding: missing value imputation before or after encoding

I want to perform a target encoding for my categorical features although I am not sure when to perform the data imputation if any of them has missing values. Let's say I have a few continuous features,...
0
votes
1answer
14 views

preparing free text column for regression

I have a column X which contains occupation/profession as an independent variable as free text, which is very much correlated with a continuous dependent variable. What techniques do you usually use ...
2
votes
1answer
131 views

Why feature transformation is needed in machine learning & statistics? Doesn't it affect the “interaction” between features?

Before feeding machine learning models, we can do data transformation and feature scaling depending on data distribution. For example, if a column is skewed, we can use Box-Cox transformation to ...
0
votes
0answers
17 views

Multiclass classification with a balanced dataset and one high-priority label

I have a balanced dataset for a multi-class classification problem with one high-priority label (this ought to be classified properly at all costs). How do I go about creating a workflow for this ...
0
votes
0answers
22 views

How to perform feature scaling on noise removel process?

i'm working on dataset contain machinery sensor data. each column(feature) represent different sensor data(pressure, temperature, speed, etc) of the machine part. here task is to predict normal ...
1
vote
0answers
53 views

Longitudinal Data with Equal Outcomes Within Individual Samples

I need to prepare some data for plugging into a predictive model. The data is in tidy format, but it comes from an audit table, i.e. every change made to a record is recorded and stored as a separate ...
0
votes
1answer
34 views

Using outlier records as a feature in model building

I am exploring the Big Mart Sales III dataset and trying to understand if using outlier rows to build a feature for predictive modeling is a sound and correct approach. This is how I have proceeded ...
0
votes
0answers
80 views

How to use seasonal features in time series regression with models such as xgboost?

I have a hard time understanding how one can create seasonal indices such as a yearly mean or (x - yearly mean(x)) and use them as predictors for monthly n horizon forecast. For example: I want to ...
1
vote
1answer
68 views

When using linear function approximation how (and why) should I incorporate the actions into the feature vector?

When reading R. Sutton: Reinforcement Learning - An Introduction (2nd edition), in chapter 10.1 Episodic Semi-gradient Control, the Mountain Car problem is mentioned and as an example it is solved ...
1
vote
1answer
12 views

Good way to use word similarity as a feature in supervised ML on text

I have a pretty low N data set of small sentences tagged with a label. I would like to create a classifier on this dataset. The word choice is not very variable since the domain is pretty specific. ...