Questions tagged [classification]
An instance of supervised learning that identifies the category or categories which a new instance of dataset belongs.
3,285
questions
0
votes
1
answer
17
views
Everything is classified as background by segmentation model
I am training a U-NET model for medical image segmentation. Problem is that the binary masks that im using to train the model mostly consist of background pixels and a very small region of the whole ...
- 51
0
votes
1
answer
18
views
How can I improve xgboost classifier if overfitting start from the initial epochs?
I am training a XGBoost multi-class classifier, but got very bad result. The train/val leaning multi-class logloss curve showed that overfitting started from the early epochs. What directions can I ...
- 101
1
vote
0
answers
20
views
How to apply CalibratedClassifierCV in external validation of a Random Forest model
I have a model trained on my data. I used joblib to get the model and shared with other teams to evaluate the performance of the model on their data. One of the team came back and said that the models ...
- 21
0
votes
0
answers
15
views
Training Data for Duplicate Detection: Allow External Information?
We have collected metadata of scientific publications (in a bilingual English-French context) from several international platforms (OpenAlex, Scopus) and French platforms (Hal, Idref, etc.). Many ...
- 1
0
votes
0
answers
16
views
Cross entropy loss for multi classification problem
I am handling a multi-class classification problem, with label in the following form [1333201000]
and the logit output of the model is in the form
([[ 0.4523, 0.0198, -0.1911, -0.0036],
[ 0.4917, 0....
- 1
0
votes
0
answers
20
views
Improving Recall and Precision of the Minority Class with XGBoost to Maximize Profits in Unbalanced Data
The company is interested in identifying profitable customers who are likely to purchase a ticket when given a promotional offer. My goal is to build a model to predict whether a customer will buy a ...
- 1
0
votes
0
answers
23
views
How to choose thresholds to discretize target for binary classification
My group is using logistic regression to investigate the most predictive features in a dataset. Our target variable is actually a continuous variable that we discretized using two cutoff thresholds (...
0
votes
0
answers
8
views
I wrote a classifier in C++ but it doesn't learn and returns the same loss each epoch
I wrote a classifier in C++ and tried training it on MNIST set but it doesn't learn. Because I am using log loss, it returns a loss of -ln(1/10) basically random chance. I tried tinkering with my ...
0
votes
0
answers
6
views
Classification for multi row observation: Long format to Wide format always efficient?
I have a table of observations, or rather 'grouped' observations (that spans more than a row), where each group represents a deal, and each row representing a product. But the prediction is to be done ...
- 143
1
vote
1
answer
73
views
How do I compute and plot Bias and Variance of a classifier in Python?
I'm new to Machine Learning and I understand bias and variance in theory but I can't seem to find a single source that explains how bias or variance can be computed. I'd like to do it in Python and ...
- 113
0
votes
1
answer
12
views
How to aggregate classes for higher overall accuracy?
I have trained a classifier on a dataset that comprises a large number of classes. Some classes are easy to predict, whereas others are frequently misclassified.
I would like to aggregate the classes ...
- 681
0
votes
1
answer
25
views
Fixing class imbalance vs Over-detecting in test data
In my experiences, binary classifiers tend do better in terms of F1 scores when the class imbalance is at least reduced. However, this leads to over-predicting in the test data.
(Thought) Example: If ...
- 131
3
votes
2
answers
464
views
Preserving / fixing class imbalance
Suppose that I have 2 collection $A$ and $B$ of unlabeled animals that are either dogs or cats.
The dogs in $A$ and the dogs in $B$ are not necessarily identical, other than the fact that they are ...
- 131
0
votes
0
answers
23
views
Fine-tuning pretrained model on 2 tasks with 2 labeled dataset
I am having difficulty using BERT for a sentiment analysis task that handles both aspect-based sentiment analysis (ABSA) and comment sentiment analysis. I know that using two separate classification ...
- 1
2
votes
1
answer
32
views
Grouping similar classes to improve accuracy, whilst maximising the number of classes
Suppose I have a large number of distinct classes, some of which are related.
My model has high classification accuracy for some classes, whilst other classes are hard to predict.
How could I group ...
- 681