Questions tagged [machine-learning]
Machine Learning is a subfield of computer science that draws on elements from algorithmic analysis, computational statistics, mathematics, optimization, etc. It is mainly concerned with the use of data to construct models that have high predictive/forecasting ability. Topics include modeling building, applications, theory, etc.
11,316
questions
0
votes
0
answers
12
views
L2 regularisation included in Validation Loss is counter intuitive?
I have been trying to tune hyperparameters for a neural network - I noticed the validation data loss for tensorflow in particular includes the L2 regularisation loss as a measure of the total loss.
...
- 101
1
vote
1
answer
42
views
Deep learning model produces very different results when classifying the same samples
I'm trying to design a simple deep learning application for biometric system verification, but every time I run the application I get very different results and I can't figure out why. I don't use ...
- 11
1
vote
0
answers
28
views
Is there a list of all the incorrectly labelled MNIST images along with their correct labels?
It seems that its well known that the MNIST handwritten digit dataset contains quite a few examples where the labels are clearly wrong and correspond to the wrong digit, some examples of people ...
- 11
0
votes
1
answer
16
views
How to represent facial features from video and classify high/low personality traits from facial features?
The dataset has 3-minute 30fps video conversations (no audio) of 150 extroverted and 150 introverted individuals. The goal is to classify them as "introverts" or "extroverts" based ...
0
votes
1
answer
36
views
Identification of inexactly-recurring material in time series stream
I am working on a personal project involving the analysis of a stream of audio data and the identification of (non-verbatim) repeated subsequences.
My research on time series has so far lead me to ...
- 101
0
votes
0
answers
33
views
Detecting cropped features/objects in an image
I hope someone can point me in the right direction. (Cross posting from the main SO flow page)
I have some images and I'm trying to write something in Python to detect if the objects in the image have ...
- 101
1
vote
1
answer
61
views
Why is T5 often used in text-to-data for text prompt encoders?
In the text-to-data(music, image, audio, etc.) generative AI field, one method of encoding text prompts is to use pre-trained language models. Such an approach was used in research on Moûsai [1] and ...
- 13
0
votes
0
answers
17
views
Text and Checkmarks Extraction from an Image
I am working on a project where in I have a filled form which is a safety inspection checklist and I am processing it through AWS Textract.
I am able to fetch text, layout, tables, signatures but ...
- 1
0
votes
0
answers
22
views
How to Fix Dimension Issues of features and classes from a Multilabel Classification dataset in getting the Out-of-Bag Error of a Random Forest?
I have created a multilabel classification dataset using make_multilabel_classification from scikit learn:
...
0
votes
1
answer
57
views
Prediction intervals for future timestamps - out-of-sample
I've created a model for out-of-sample forecasting that uses multistep recursive strategy to reduce my problem to regression, the predictions are sufficient but I was wondering if there is any ...
- 1
1
vote
0
answers
61
views
How does fine-tuning work in question answering for custom documents
I am trying to build a Q&A bot for which I have a bunch of documents like articles (specific domain).
I understand I can create a Retrieval-Augmented Generation (RAG) system for this, but I want to ...
- 23
0
votes
0
answers
114
views
Convert specific domain knowledge text to a knowledge graph
As part of this semester assignment , I'm working on a project that aims to to represent the knowledge in "PMBOK 6th edition, section 11: Project Risk Management (page 395 -> 458)" and the knowledge ...
0
votes
0
answers
20
views
Why is the feature direction chosen in the direction associated with largest eigenvalue of $Σ_T$ in case of more than two classes?
Why is the feature direction chosen in the direction associated with largest eigenvalue of $Σ_T$ in case of more than two classes? Please see the following.
- 101
0
votes
0
answers
45
views
Accuracy difference between 1-channel grayscale and 3-channel grayscale detection model
I have found no similar questions to this online, or answers for that matter.
I am using cameras that output a grayscale image, which I feed into a Yolov8 object detection model (Specifically yolov8m-...
0
votes
0
answers
32
views
Regresson on 3d tabular data
There is a dataset, where each Y depends on separate table :
Each table consists of the same set of columns, but the number of rows may vary.
Features in tables are both numerical and categorical (...
- 13
0
votes
0
answers
22
views
Building an application to group/label browser tabs
I'm trying to develop an application where users can click a button and all of the open tabs in their browser will be placed into tab groupings based on similarity of the tab. Microsoft Edge has a ...
- 1
0
votes
0
answers
63
views
How to deal with "Could not broadcast input array from shape (1141,2) into shape (1141,)" to get Out-of-Bag error while using Random Forest
I have a dataset that consists of 171 features and 39 labels. I captured both features and labels of the dataset through slicing:
...
0
votes
0
answers
32
views
Torchvision Faster-RCNN, modified loss function
I'm trying to solve a problem of table detection in spreadsheets in Excels. I've came across this paper, which suggests to use modified version of Faster RCNN to do object detection on the ...
- 101
0
votes
2
answers
19
views
Looking for suggestions on the model/algorithm for 2D row labeling
As an experiment, I'd like to train a model to label a few rows of data on a 2D tensor. F.e. on a black and white image, label the "darkest" and "lightest" row.
Is this a task for ...
- 103
0
votes
0
answers
27
views
Incorporate a new feature or Post-process
Briefly, I am training a model using XGBoost to predict future quantity for the factory to produce. Basic features currently in use are date time features, categories, holiday (binary). I have just ...
- 1
0
votes
0
answers
23
views
In $n$ dimensions, why do we need at least $n+1$ sample points to have covariance of sample data non-singular?
In $n$ dimensions, why do we need at least $n+1$ sample points to have covariance of sample data non-singular?
0
votes
0
answers
28
views
Training Biased/Uneven Categorical Data with CatBoost, Unbalanced/Unseen Categories Handling
Summary:
I am training a discount eligibility model where the dataset represents historical data for products where people availed discounts based on simple features like product category, discount ...
- 101
0
votes
0
answers
11
views
RecSys model performance stalling at 47% AUC and F1-Score. Is the problem due to ratio of users to items in my dataset?
I'm having trouble with making my validation metrics go down for the binary_crossentropy and go up for the F1-score and AUC. I've tried tuning my hyper parameters such as the number of latent features ...
0
votes
1
answer
17
views
Unclear point in outlier construction based on an heavy-tail distribution
The following is an example from a book (An Introduction to Pattern Recognition and Machine Learning by P. Fieguth, page 417) on random variables and random vectors. Please consider the following.
...
1
vote
1
answer
32
views
How in the heck should I tackle this classification problem? I'm not even sure if it's classification or regression
So, I'm currently a third year student in electrical engineering and I'm currently enrolled in a Mathematical Modelling and Machine Learning class and we're currently tasked to classify or use ...
- 11
0
votes
1
answer
42
views
Is there a way to focus mainly on high precision when fitting a tree model?
I have a dataset with 95% false and 5% true labels, some 200000 samples overall, I'm fitting a LightGBM model. I mainly need to focus on high precision and have low number of false positives, I don't ...
- 1
0
votes
1
answer
58
views
Unclear points on projection type and selection of distance metric in feature extraction for a set of scenarios
The following is an example from a book (An Introduction to Pattern Recognition and Machine Learning by P. Fieguth, page 85) on feature extraction and selection. Please consider the following figure.
...
0
votes
0
answers
19
views
Simple classification problem between real and fake data always gives zero loss
Hey I have just started with machine learning and was trying to classify between real (label 1) and fake data (label 0). Here is the code that I have written. The ...
- 101
0
votes
1
answer
19
views
Approaches to dataset, whose elements have different size
I am working with a dataset where each elements is a square table of size m-by-n, where m (the number of rows) is the same for all the data points, while n (the number of columns) varies from tens to ...
- 101
0
votes
0
answers
26
views
Plot binary logloss of LightGBM
I implemented Gridsearch for a LightGBM, predicting a binary outcome with 33 features.
...
- 13
0
votes
0
answers
10
views
Is Kernel Density Estimation actually the continuous/curved version of a histogram?
I have been studying ARIMA for a bit of time now, and stumbled upon the plot_diagnostics function.
Among others, it plots out the Histogram & KDE on the same plot. As I did not know anything about ...
0
votes
0
answers
8
views
how to derive this negative sampling approximation for graphs?
I am watching CS224W video (https://www.youtube.com/watch?v=Xv0wRy66Big&list=PLoROMvodv4rPLKxIpqhjhPgdQy7imNkDn&index=8) and have some issue understanding the following approximation:
Does ...
- 161
0
votes
1
answer
21
views
Game Score Projection Regression Model Accuracy Improvement?
I'm creating a regression model to project scores of a basketball game. Here are my training results:
loss: 9.4327 - val_loss: 9.4029
I've tried adding more features, messed around with the model (...
0
votes
3
answers
75
views
Reduce false positives having imbalanced data
I'm using a DNN-48 having the following scenario:
Features: 8 (48 at the end because I generate conditional sequences of 6 elements each)
Classes: Y=0 (90%), Y=1 (10%)
Precision and recall are good ...
- 3
1
vote
1
answer
49
views
Interpreting ROC Scores for non-binary data
I have been handed a random-forest based machine learning model that is predicting a rate rather than a binary category (e.g. the rate of road quality deterioration in a city for planning repaving ...
- 11
0
votes
0
answers
33
views
XGBoost Classifier + Isotonic Regression leading to worse probability accuracy
I'm testing out an XGBoost Classifier with the goal of using the probabilities it predicts in production. I know that tree based model probabilities are often not calibrated well so I decided to test ...
- 1
0
votes
0
answers
60
views
Prediction in multiclass classification
Context: I need to make an multiclass classification to predict what type of sentence(law) the case will have in the end.
Data: I Have several columns to predict the case:client, cause of action, ...
- 1
0
votes
0
answers
34
views
Training loss is much higher than validation loss
I am trying to train a neural network with 2 hidden layers to perform a multi class classification of 3 different classes. There is a huge imbalance to the classes, with the distribution being around ...
0
votes
0
answers
15
views
0
votes
1
answer
21
views
How to pass a Dataframe as train dataframe and another dataframe as Validation to GridSearchCV
I'm a programmer who tries to find he's way into ML world. so the Question might be basic.
i have data from years 2010-2019. Now i'm trying to test different parameters on gradient boosting regression ...
- 101
0
votes
1
answer
33
views
Is this Dataset somehow skewed?
I am working on a dataset that has 100K points, it's about Customer churn. So I don't know whether this dataset is skewed, incomplete or what. I tried doing some feature engineering on it but couldn't ...
0
votes
1
answer
28
views
Why did both my precision and recall (hence also F1) improve dramatically but the accuracy lowered?
I split all the audio recordings in my small dataset into short clips, thereby "created" more samples. All of precision, recall and F1 almost doubled (from around 0.35 to around 0.65), but ...
0
votes
0
answers
86
views
Parsing response from llama2
I want to extract phone numbers from a given text and i am prompting a llama2 model for that ..I want the output in form of a list but i am getting unnecessary output like sure here are the phone ...
0
votes
0
answers
15
views
If my logistic regression model is performing well, does it matter if my features don't pass the Box Tidwell Test?
I've built a logistic regression model for binary classification with a high F1 score, but when I run Box-Tidwell tests on continuous independent features/predictive variables, I find non-linearities ...
0
votes
0
answers
20
views
Can you please provide one fully solved gradient boosting regression numerical example (not python code) [duplicate]
Can you please provide one fully solved gradient boost regression numerical example (not Python code).
0
votes
1
answer
91
views
Best metrics to evaluate the performance of a regression model?
I've just started with machine learning and I have a lot to learn but one of the recent problems I'm facing is evaluating the performance of a regression model. I know about MSE, RMSE, MAE ...
0
votes
0
answers
138
views
0
votes
0
answers
70
views
How does supervised fine-tuning work in InstructGPT?
See Figure 2 from the InstructGPT paper:
I want to know how Step 1 works. Here is one possible algorithm.
Pass the prompt through the model, and compute the negative log of the probability of the ...
- 135
0
votes
0
answers
17
views
Collaborative Filtering using ALS
We are trying to use collaborative filtering mechanism for recommendations, implicit data, based on users are navigating to. Trying ALS (using Spark) which makes sense here. All fine.
Now the model ...
0
votes
2
answers
43
views
Is SVM is a good choice for large dataset?
With my limited knowledge of SVM, I am following a tutorial on YouTube to create an End-to-End multi-class ML model . There the person is using SVM on a dataset with 9 images dataset, but the dataset ...