Skip to main content

All Questions

Filter by
Sorted by
Tagged with
4 votes
1 answer
643 views

Rule based label - random split vs time-based split

We have a dataset of 977 records (77:23 class ratio) where we try to predict a binary outcome using random forests and neural networks. whether supplier met the target or not. However, we didn't have ...
The Great's user avatar
  • 3,342
3 votes
2 answers
684 views

test & train for very very small data

I have just 25 observations. I'm not sure would it possible to test & train the data. For example 15 observations for train and 10 observations for test set. 15 observations is so small for ...
Leila ali's user avatar
  • 179
3 votes
1 answer
2k views

Why does error rate of kNN increase when k approaches size of training set?

I've been experimenting with the effect that different values of k have on the generalisation error of kNN classifier, and I've gotten some unexpected results towards the end when k approaches the ...
namiyousef's user avatar
2 votes
1 answer
3k views

random split vs time based split of train and test data

I have been working on binary classification problem using algorithms such as Random Forest, Boosting methods, neural networks and logistic regression. I have data from Jan 2017 to Jan 2022. We wish ...
The Great's user avatar
  • 3,342
2 votes
1 answer
36 views

Model Stacking Train Test Split Methdods

I am trying to validate my processes in terms of how I am engaging in model stacking for binary classification. Say I have two models as my base models, models A and B both with different classifiers ...
user54565's user avatar
1 vote
3 answers
138 views

How to design cross-validation and testing scheme when N is small?

I have a binary classification problem with 60 samples (N=60). 40 are responders (+) and 20 are non-responders (-) to a drug. There will be ~20 measured features (p=20) per sample with which to make ...
zachursinus's user avatar
1 vote
1 answer
48 views

CV score vastly different from Train-Test score

I'm working on a multi-class classification task. I'm currently trying to tune a LGB model but have encountered a behavior that I do not understand. First, my data is from 1996 to 2015 so I split my ...
jauyjad's user avatar
  • 11
1 vote
0 answers
36 views

Choosing a suitable sample size for a Random Forest Model

I know that this isn't always a straight forward question to answer, but I am working on a provincial wide wetland classification model that has $7$ classes and $32$ or so explanatory variables. In my ...
John Gallop's user avatar
1 vote
0 answers
19 views

Which one to choose: subject-wise or record-wise

I want to do some classifications on open access "Activity Recognition Using Wearable Physiological Measurements" data set. In this dataset, there are 40 subjects whose Electrocardiogram (...
Mina's user avatar
  • 111
1 vote
0 answers
41 views

How do I perform a train-validation split on data with class imbalance such that the class imbalance ratio is preserved?

My data has class imbalance-- that is, some classes have significantly fewer training samples than the others. I want to perform a train-validation split in such as way that the class ratios are ...
skinnybb's user avatar
1 vote
1 answer
52 views

Train/dev/test split with limited and skewed positive labels

(Because of the sensitive nature of the actual project, I am using an analogy here. I hope it's clear, if not, please let me know!) My goal is to classify images as cats or dogs (binary classification)...
stinodego's user avatar
  • 113
1 vote
0 answers
100 views

Splitting Data Into Training and Test set [closed]

I am trying to split a data set into training and test set with these codes ...
M. syed's user avatar
  • 11
1 vote
0 answers
48 views

Classifier score is higher on test set than on training set. Is this an error?

I am currently getting used to scikit-learn and I trained a simple logistic regression model on the iris dataset. One interesting thing I noticed was when I looked at the classifier.score, namely: <...
Bix's user avatar
  • 11
0 votes
0 answers
24 views

Model Stacking - Out of Fold Procedures

I am attempting to use a model stacking procedure where I am using a time-series split on a set of data I have (around 5000 entries). The goal is binary classification. After obtaining hyper ...
user54565's user avatar
0 votes
0 answers
15 views

Is GroupKFold needed if some samples have some of their feature values equal?

I am given a dataset $D$ of 10k enzyme-substrate complexes having a lock-key relationship, with each sample (complex) being characterized by enzyme features $x_e$ and substrate features $x_s$. That is,...
Antonios Sarikas's user avatar
0 votes
0 answers
39 views

Is it normal to have a sharp increase in validation error when using 10% of the data instead of something like 20-30%?

Scenario: I'm training a relatively simple neural network to classify pairs of tabular datapoints (~150k), lets say drugs and diseases, whether they are related (positive) or not (negative). As I only ...
oliver.c's user avatar
  • 185
0 votes
0 answers
84 views

Statistical comparison of two probabilistic classifiers

TL;DR There are well-known tests to compare classifiers. How can we generalize to classifiers with random training steps? I am comparing two classification algorithms (A and B). I can see that ...
independentvariable's user avatar