Questions tagged [feature-engineering]
the process of using domain knowledge of the data to create features that improve machine learning algorithms
0
votes
1answer
24 views
Feature Importance Scores Python
I have a dataset having 7 attributes viz., time, C1, ... C7 pertaining to earth quake reports where each column/attribute represents a certain aspect of damage viz., power, sewer_and_water, ...
1
vote
1answer
19 views
Hand-crafted decision tree inspired from learned decision tree
Goal of this question:
As I am the only 'machine learning guy' in our group, I wanted to get an outsiders view, that is a sanity check if what I am doing adheres at least to 'decent practices' in ...
0
votes
0answers
7 views
How to do feature engineering to the stripplot where the target, `tradeMoney`, has obviously lower than 5000 when 'rentType' is 'shared_rent'?
I am dealing with a house prediction problem.
When I am doing EDAs I find the such stripplot() as follows:
...
0
votes
0answers
20 views
Difference between three equivalent ResNeXt blocks
https://arxiv.org/abs/1611.05431
I have been reading this article and have a question about the following three equivalent ResNeXt blocks. In the article, it says
Under this simplified case, ...
0
votes
0answers
10 views
We know the subspace generated from the data instances, but we cannot constitute the origin space
I was wondering, what if we know the subspace generated F from the data instances, but we cannot constitute the origin space E that can be in higher dimension, and can easily lead us to the true join ...
0
votes
0answers
24 views
How to choose feature which is used to fill another feature's missng values?
I am dealing with a house prediction problem. However, it has about 10% missing values in buildYear which is one of the most important features. I tried filling ...
0
votes
1answer
23 views
how to deal with two high correlations feature which both has a low correlation with target
I am doing a prediction of house trade money.
Here is the correlation matrix of features whose correlations are larger than 0.3 as follows:
...
0
votes
0answers
12 views
When to create ranges for numeric feature?
For example, in the Titanic Dataset, I'm trying to deal with the numeric datasets, FamilySize and Ticket (Ticket Price).
From the many solutions I've seen, a lot of people create ranges for ...
0
votes
0answers
21 views
convert significant features to a set of rules or information
Is there any way to set up some rules from features in a classification model?
Assume that we want to classify an employee as someone who will be terminated or not. We found that average hourly pay ...
0
votes
0answers
16 views
Predicting U.S. suicide count based on a set of inputs
I'm trying to design a model (or multiple) that can predict the number of U.S. suicides for a future year, based on a few inputs--"age", "sex", "population" (of the age/sex), and "gdp_per_year". I'm ...
2
votes
1answer
23 views
How to choose an optimal threshold for binary discretization
We know that we usually do discretizations to continuous features to remove extra information and unwanted regularities, which makes the model robust and well-predicted.
But I am wondering except ...
0
votes
2answers
35 views
How to encode a job description for machine learning
I'm working on a sample project and one of the features is the job description of a person (categorical, for example: blue-collar, retired, unknown, unemployed, student, etc.). Since in the future ...
1
vote
0answers
32 views
What are the audio features to best describe a music?
I'm working on the content-based filtering part of a recommender system for an audio streaming project.
I firstly used the k-mean algorithm with music genres and one-hot encoding to classify musics ...
0
votes
1answer
13 views
How to combine features with different temporal scale in machine learning
We have various types of data features with different temporal scale. For example, some of them describe the state per second while others may describe the state per day or per month from another ...
2
votes
2answers
78 views
Combining Latitude/Longitude position into single feature
I have been playing with 2 dimensional machine learning using pandas (Trying to do something like this: https://github.com/freeman-lab/spark-ml-streaming), and I'd like to combine Lat/Long into a ...
4
votes
1answer
40 views
Model for Differing Number of Rows per Observation
Looking to build a response model (click or no click) on marketing data which displays varying number of offers to a person. I don't want to model which offer they click but do they click any of the ...
1
vote
0answers
11 views
if new feature downgrade the score for xgboost what do I have to look at?
let say I'm predicting the housing price of Boston(kaggle).
if I got some score x
then I added new feature y_K
if this new feature drop the score.
what is wrong with this feature and what do I ...
0
votes
0answers
48 views
method and dataset for credit card fraud detection
I am trying to create a machine learning model to detect credit card fraud (In our definition, fraud means chargeback). I am kinda stuck now with the dataset that I have. I don't have information on ...
2
votes
5answers
368 views
Categorical vs continuous feature selection/engineering
I'm working with a dataset with a number of potential predictors like :
Age : continuous
Number of children : discrete and numerical
Marital Situation : Categorical ( Married/Single/Divorced.. )
...
5
votes
1answer
135 views
Feature engineering suggestion required
I am having a problem during feature engineering. Looking for some suggestions. Problem statement: I have usage data of multiple customers for 3 days. Some have just 1 day usage some 2 and some 3. ...
1
vote
0answers
15 views
Aggregating target-encoded array-like categorical features?
I am trying find commonly used techniques when dealing with high cardinality multi-valued categorical variables for machine learning classification algorithms.
One-hot encoding leads to very high ...
1
vote
0answers
19 views
How to incorporate an attribute that only exists in some observations?
In a binary classification problem, some of my observations have an event that occurs. I can, obviously, add a 1/0 flag if the event occurs ("event_occurred" in the data below). However, my intuition ...
2
votes
1answer
42 views
Do I need to engineer lagged features when creating an LSTM for time series forecasting?
Long short-term memory networks are fairly complicated and I haven't completely wrapped my head around them.
It seems to me like the big gain in LSTMs for time series forecasting is the lacking ...
0
votes
2answers
33 views
Pros and cons of using the zscore of a dataset before normalizing it during feature engineering?
Normalization is a common feature engineering technique. However, this post used standardize(zscore) on the dataset before normalizing it.
I think that would result in losing some of the information ...
2
votes
2answers
36 views
Creating a metric based on some features
I want to create a new metric based on some features but dont know how to start. I basically want to create a "job satisfaction level" metric based on some features. The features could be work hours, ...
0
votes
1answer
116 views
5
votes
3answers
243 views
Regression vs Random Forest - Combination of features
I had a discussion with a friend and we were talking about the advantages of random forest over linear regression.
At some point, my friend said that one of the advantages of the random forest over ...
0
votes
1answer
37 views
Machine learning algorithm that can use many instances to predict 1 continuous outcome per person
I am trying to use movements identified from accelerometers during sleep to predict gait speed (continuous). I am trying to figure out what the best machine learning algorithms/ feature extraction ...
1
vote
1answer
26 views
is it bad to have many different measurements for the same target variable?
I'm working on a dataset that has repeated measurements for the same target variable.
When I don't change anything and create model, cross validation overfits with 0.99 score but in testset it gives ...
2
votes
0answers
31 views
How to deal with a potencially multiple categorical variable
I'm build a model that has, as inputs, some categorical variables. I had already dealt with this sort of data before, and applied different techniques as creation of dummy variables and factor scoring....
0
votes
1answer
38 views
How exactly do I extract the important features from strings for machine learning?
Forgive me for my ignorance. Linked below is an image of my dataset with 1000 tuples.
https://i.stack.imgur.com/WHIlx.png
I have the following questions
(1) How exactly do I go about extracting ...
1
vote
2answers
44 views
How exactly do I go about extracting features from timestamps for machine learning? [closed]
My dataset has a timestamp column with the following format: 06/24/18 0:56
How exactly do I convert this information into features that can be used for classification algorithms like logistic ...
3
votes
2answers
88 views
Manual feature engineering based on the output
So, I'm working on a ML model that would have as potential predictors : age , a code for his city , his social status ( married / single and so on ) , number of his children and the output signed ...
1
vote
0answers
57 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,...
2
votes
1answer
69 views
Why is Reward Engineering considered “bad practice” in RL?
Reward engineering is an important part of supervised learning:
Coming up with features is difficult, time-consuming, requires expert knowledge. "Applied machine learning" is basically feature ...
0
votes
0answers
12 views
Add features from a timeserie target to train set
I have a one year train set, which is a combination of a sequence of time ordered images, while the target is a continuous variable.
The source of the problem is that data-set is very small, ...
1
vote
1answer
73 views
Multiclass classification in a balanced dataset with one high-priority label
I have a balanced dataset for a multiclass 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
1answer
14 views
Inverse Binary Feature
I am feeding a binary value into my NN which represents whether the given example is a public holiday or not.
Is there a difference between assigning a 0 to public holidays and 1 to all other days or ...
2
votes
2answers
71 views
How to Work with Imbalanced Data
I am building a binary classifier from a set of feature vectors some of which are categorical like Yes or No (two options). I am replacing them with 1 and 0 and since there is strong imbalance between ...
1
vote
1answer
201 views
Python Time series: extracting features on a rolling window basis
I have a long univariate time series, and before performing some machine learning models with it, I want to extract as many features as I can from the time series on a rolling-window basis.
As a ...
1
vote
0answers
33 views
Adding Fourier transform features to data
I'm working on some timeseries data which after visualising seems to be periodic(repeating at some interval), So I planned on finding the Fourier transform of the entire and pick the top n amplitudes ...
1
vote
1answer
23 views
Feature Engineering Lists\Vectors as values in dataframe
Let's say I have a dataframe where some of the columns have lists of strings as values. I would like to use ML Algorithms on this dataframe.
In this case, I can:
I could add many columns of 1's and ...
1
vote
0answers
27 views
Feature encoding for multiple JSON objects
I have a dataset, where a particular feature is a collection of many JSON objects for a single feature.
...
2
votes
0answers
28 views
Measure the “aggregate preference” of points on a 2D plane
Consider some points on a 2D plane, we know corrdinates of all
points, how to measure the "aggregate preference" of them.
I don't know what is the exact terminology of
"aggregate preference".
What I ...
1
vote
1answer
26 views
Should I expect major performance improvements by scaling my features?
I'm trying to decide whether I should scale my features & responses for training, and I'm in a situation where I can't just try both scaling and not scaling.
My features currently have an std ...
1
vote
1answer
32 views
Problem with important feature having a lot of missing value
I am facing a dilemma with a project of mine. One of the variables (numerical) doesn't have enough data i,e almost 99% data are missing. However, upon talking to the domain experts, it appears that ...
0
votes
1answer
75 views
ML: How to think feature selection?
What is the basic philosophy behind feature selection and modelling? How do you actually start? Could you please share your real (practical) inputs?
Bit of background:
I am actually trying to analyse ...
0
votes
0answers
7 views
Provide optional confidence level as an input to the neural network
I have a name, gender labeled dataset and I know the frequency of particular name can occurred in the dataset. I want to develop a neural network which predict gender when given the name as an input. ...
1
vote
1answer
22 views
Why feature crosses may work better than having them as individual features?
On Google ML Crash Course it is said the following:
If we build a feature cross from both these features:
[behavior type X time of day]
then we'll end up with vastly more
predictive ...
0
votes
0answers
24 views
Feature selection through Random Forest and Principal Component Analysis
I am working on a binary classification problem and I have 870 numeric independent features to start with. I tried PCA on input features and picked top 200 variables corresponding to first 10 ...
