Questions tagged [machine-learning]

Machine learning algorithms build a model of the training data. The term "machine learning" is vaguely defined; it includes what is also called statistical learning, reinforcement learning, unsupervised learning, etc. ALWAYS ADD A MORE SPECIFIC TAG.

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

Can a convolutional neural network be used to classify an image by the number of X in an image?

Say, for example, I were to make a leukemia classifier. The disease is characterized by an excess of leukocytes. However, I'm not sure if a CNN is the best method for this sort of classification ...
1
vote
0answers
4 views

On quantifying the amount of information per example provided to the model in Supervised vs Self-supervised learning

I've seen Yann Lecun in his self-supervised learning talks talking about how traditional supervised learning (Classification setting) by attributing a class out of N classes to each example the ...
1
vote
0answers
24 views

Why does changing my training set by ONE point dramatically affect CV error?

I'm working on a time-series, binary classification dataset, where I'm doing cross validation as a moving window as in the diagram below: So I'll use the first three "shifts" as cross ...
2
votes
1answer
28 views

Square root transformation of Poisson process. How $\small Var[\sqrt{P(\lambda)}] \approx \frac{1}{4}$

I am working on Kaggle Neural data challenge. I am trying to understand the transformation applied on the neural spiking data. A number of spikes given a stimulus are Poisson distributed as $$Y_i \...
2
votes
1answer
22 views

When to use accuracy_score with train_test_split VS. cross_val_score with kfolds?

I want to measure how well my predictive model performs so I can report the percentage for my project, but I'm not sure which of the scores to report, the one given when I used kfolds and ...
0
votes
0answers
13 views

Add k-means layer to Keras model

I'm using a GAN to generate pixel-art images. The structure follow the Tensorflow tutorial on how to do GAN closely. My network outputs gradient-rich images, which look like down-scaled photos rather ...
1
vote
0answers
16 views

Transforming y and x-axis

I have a model that predicts frequencies. True frequencies are close to 1 therefore the frequencies that predicted are mostly close to 1. Sample frequencies are: ...
1
vote
1answer
19 views

Data dimension in machine learning

I am working on Ml project and I Have 4-d dataset. I wanted to use dimensionality reduction algoritm And suddenly a question made me stop Here is my dilemma Is there difference between dimension ...
5
votes
2answers
187 views

For hyperparameter tuning with cross validation, is it okay for the fold splits to be same for every hyperparameter trial?

For hyperparameter tuning (random search/ grid search/ bayesian optimization), there are many trials performed for each set of hyperparameters. To evaluate how good a set of hyperparameter is, we can ...
0
votes
0answers
10 views

Statistical test for determining difference from true measurement?

So hypothetical situation is as follows: Let's say I make a neural net that can determine the length of an object in an image in centimeters. I also have an actual length for that object. My goal: ...
0
votes
0answers
12 views

Tree's branch ends to the same leaf twice

I am in r using the DoctorVisits dataset from the AER package. I chose the column ...
0
votes
0answers
16 views

Approach(es) to discover “conditional churn propensities”?

I'm aware of a variety of approaches to discovering “stand-alone” churn probabilities. But I haven't been able to find by searching any info on “conditional” churn probabilities. Use case: I'm a ...
2
votes
1answer
49 views

What is condition of a smallest variance estimator?

I ran into a confusing problem as following: Linear regression estimator has the smallest variance among all unbiased estimators. The answer is said this fact is false, but based on LS notes and ...
2
votes
1answer
17 views

Sparse and variational Gaussian Process lower bound

I am learning about the Sparse and Variational Gaussian Process by Titsias. The explanation for Variational Sparse GP starts at around 1 hour 12 min mark. In slide 29 of the lecture, the author stats ...
0
votes
0answers
6 views

PCA of self generated data using orthogonal components itself, gives back inconsistent components

I was trying to study python API in sklearn for PCA to test if it can recover back the orthogonal vectors to generate data. Steps Used: Generate 3-D random orthogonal vectors. I take first two ...
0
votes
0answers
8 views

Why do autoencoders with bias terms converge to a constant mapping of the mean?

I'm currently reading the paper Deep One-Class Classification by Lukas Ruff et al, and my questions are about footnote 1 at the bottom of page 5 of the paper, and about proposition 2 in section 3.3 on ...
4
votes
1answer
193 views

What assumptions about probability do different models make?

If we throw a dice n times and get an empirical, discrete probability distribution, one could say that it approximates the real underlying distribution. When we build a machine learning model like ...
0
votes
0answers
7 views

What is the best programmatic way to find the best ensemble model in python, as to which models are best suited to which portions of the data

Generally in ensemble modelling as dataset is being segregated into multiple portions where each portion is being trained on a particular model, what is the programmatic way to determine which model ...
0
votes
0answers
25 views
+50

Find optimal training dataset after concept drift

There are many strategies how to detect a concept drift or model drift, like when there was a major change in the underlying process so that the model becomes invalid. It can be an abrupt change or it ...
0
votes
0answers
7 views

Large Dataset vs small Splitted Dataset for XGBoost [closed]

I have a Promotions Dataset consisting of around 500,000 records for each retailer. Owing to the nature of the project, I have chosen XGBoost Algorithm. Now, what would be a good approach:- a) Train ...
0
votes
0answers
18 views

After Deep Learning Hyperparam tuning, what adjustments should be made when dataset size is scaled up?

I'm dealing with a fully connected NN, and I'm wondering if there are any rules of thumb for adjusting hyperparameters for changes to dataset size. For example, if I increase number of obs by 20%, ...
0
votes
0answers
11 views

Feature engineering and longitudinal data

I need some advice for my feature engineering. Suppose I have 90 days follow-up data. on 12 patients and I have the vital status of the patients at the end of these 90 days (deceased=1, alive=0) ...
1
vote
1answer
14 views

How many emails would I need to train a good text extraction model?

I'm looking to train a model that will identify product names in an email that a user has bought. The end result would be something very much like named entity extraction, except this should correctly ...
0
votes
1answer
10 views

How I can use 'time duration' as a feature in ML models? [closed]

I had two datetime columns, I subtracted them and get the duration period, how I can further use this data as a feature in ML models.
3
votes
1answer
24 views

correcting for extremely downsampled data: keras class_weight is hurting my model

I have an extremely imbalanced dataset (millions of times more negatives) for a binary classification NN model. I am aggressively downsampling solely for the purpose of making training time manageable,...
0
votes
0answers
24 views

Theoretically: Is there any advantage/reason on using a particular model for ML over another?

A neural network doesn't really care about the activation functions, and if we choose any activation function and a compatible loss, the model will converge into something. In a way, any model will be ...
0
votes
0answers
8 views

Out of distribution prediction, not sample

My dataset contains samples that are recorded time signals. I want to predict the start of some event in the time signal. Say I train a neural network with mean squared error as loss function to ...
0
votes
0answers
6 views

Softmax as a measure of uncertainty, not certainty?

I am aware that the softmax output of a neural network is not a good confidence measure (see Gal 2016, page 13 and 14). The reasoning behind this is that they are too over confident when they actually ...
0
votes
0answers
28 views
+100

How to calculate uncertainty for predictions coming from cascade of models?

I have developed a bunch of models to predict house prices. It is a 3 fold process: I fit a gbm (first_model) and get the first prediction (first_pred), there are some sub-models (simple lineer ...
0
votes
0answers
15 views

Using Random Forest vote scores as a matching variable?

I am working on a project where we need to identify good counterfactuals / matches for a binary treatment, which is regressed against a binary outcome. The "treatment" that we seek to study ...
0
votes
0answers
29 views

Propotion test vs T-test

Proportion test - Helps to identify if there is significant difference in proportion across the samples.(i.e. the values are in proportion and denominator is in sample size of entire proportion) ...
0
votes
0answers
6 views

comparing timeseries residuals for different categories

I have a forecast model for epidemiology data (let's say covid cases). I'm trying to compare its performance across different segments e.g. males vs females: whether the model performs (in terms of ...
2
votes
0answers
31 views

Optimizing this log-likelihood

I have a HMM which emits an observation Z. The parameters of the HMM are $\boldsymbol\theta$. $$\boldsymbol\theta = {\boldsymbol{A},\boldsymbol{B},\pi}$$ Where $\boldsymbol{A}$ is the transition ...
0
votes
0answers
13 views

Understanding discriminator function

This is the function I am trying to understand is: I know that the function determines whether the image looks more realistic than fake data, or less realistic than real data, but as far as how it ...
0
votes
1answer
21 views

DNN Cannot Stop Overfitting

I am training a DNN (CNN + RNN) for a voice conversion task. Although my train loss can be very low with good performance, I believe I am experiencing massive overfitting. To overcome this, I have ...
0
votes
0answers
8 views

When would AUC fail in comparing models? [closed]

It is possible that a classifier might have 1 threshold where there is highest possible true positive rate and least possible false positive rate (and lets say that is what the application requires), ...
0
votes
0answers
14 views

Should I use LIME and SHAP on different models and observations?

I'm using LIME and SHAP to explaining features importance and to explain which features and values affected my classification model. The obvious work method is to choose the best prediction model and ...
0
votes
0answers
4 views

How to calculate and compare ML models Recall estimators from stratified sample?

Let's say that I have 2 machine learning models for a disease (binary) classification task. I would like to estimate Recall for both models and compare it (show statistically that for example Recall ...
0
votes
0answers
8 views

Where and how can I find enough data to build a chatbot? [closed]

There are so many datasets that are available on the internet such as The WikiQA corpus, OPUS, HotpotQA, RecipeQA, etc. But using public data is not enough because it is too generic. Chatbots need a ...
0
votes
1answer
23 views

Can we use Gradient Descent in the place of Ridge Regression in overfitting problem while doing linear regression problem?

What is the difference between Gradient Descent and Ridge regression? We use ridge regression for overfitting problem when the Mean Squared Error for test dataset is high. I think that we can use ...
0
votes
0answers
9 views

Do I average confusion matrix when I use cross-valiation for my classification model performance?

I am trying to build a classification model with 10-fold-cross-validation. I don't know if I have to look at the model performance by averaging the 10 confusion matrices or choose the best confusion ...
0
votes
0answers
7 views

Multi Target Techniques where Dependent Variables are Correlated

I have browser data that contains over 100 independent variables to predict customer spend. Instead of predicting total spend over a given time, let's say we want to predict the monthly spend for each ...
0
votes
0answers
10 views

Why eigen vector of a covariance matrix is the largest principle components? [duplicate]

I am self studying principle component analysis using this tutorial, I got most of the reasoning behind PCA but I don't get the intuitive reason why eigen vectors of a covariance matrix is also its ...
0
votes
0answers
8 views

Feature Engineering and prediction with R and python [closed]

I have a sequential dataset, and have 2000 rows for 300 ID. I have 20 variables (i.e in my real dataset) ...
1
vote
1answer
14 views

Consequence of application of activation function to input layer

I wasn't aware of the fact that the activation function gets applied for the first time during prop in the first hidden layer and not already in the neurons of the input layer. If I now did otherwise ...
1
vote
1answer
20 views

How can I adjust predicted probabilities after resampling?

I have a real-world problem with severe imbalanced classes. I was able to get a good AUC and balanced accuracy after the implementation of a resampling technique. Now I want to "walk over the ROC&...
0
votes
0answers
10 views

Goodness-of-fit test for conditional distribution

The chi-square goodness-of-fit test allows us to test if a data sample $(y_n)_{n=1,\ldots,N}$ agrees with some proposed model distribution $P(y)$. However, in a typical machine learning setting we ...
0
votes
0answers
8 views

How can i increase the r2 value on validation data? [closed]

I'm having a problem finding a model for my regression problem, I've tried various models with no success. I'm using 5 fold cross validation and optimizing for the r2 metric, but I get results similar ...
2
votes
1answer
31 views

Is this case possible for Decision Tree?

I am studying decision tree and I would like to know if this case is possible: We have 2 features, each does not decrease the Gini of the previous node (=> not choose), but their combination (two ...
0
votes
0answers
11 views

What is the difference between “fast-weights” and “slow-weights” in deep learning? [closed]

I'd like to know the difference between "fast-weights" and "slow-weights" in deep learning, and how the algorithm work. Are these concepts only used in RNN based models? I read the ...

1
2 3 4 5
327