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
0 answers
2 views

Coefficient of correlation between differently spaced time series

I have a timeseries with input flow values in a water plant (measured every 5 minutes) and the meteorological data containing the precipitation in the last 1,3,6,12 and 24 hours (each displayed every ...
0 votes
0 answers
8 views

The multivariate Inverse-Gamma

On wikipedia they give a multivariate form, which to my understanding is used when V is known up until the scaling factor for a Normal-InverseGamma conjugacy. I tried to find a source of the ...
2 votes
1 answer
1k views

Deriving step size/learning rate in the hinge loss passive-aggressive/perceptron algorithm

Recall the perceptron algorithm: cycle through all points until convergence $\text{if }\, y^{(t)} \neq \theta^{T}x^{(t)} + \theta_0\,\{\\ \quad \theta^{(k+1)} = \theta^{k} + y^{(t)}x^{(t)}\\ \}$ ...
0 votes
1 answer
171 views

Machine Learning Models for Classification with Categorical Variables

To start, I'd like to say I have very little experience in machine learning, or statistics/computer science in general. What I am interested in is a list of models I can use to classify a binary ...
1 vote
0 answers
9 views

Selecting a test to prove that the observed changes in the performance of two machine learning models are statistically significant

I have developed two machine learning models which I evaluated with two different datasets. My initial hypothesis was that their performance would be higher in dataset 2 as compared to dataset 1. ...
0 votes
0 answers
19 views

Without encoding, how can we solve high cardinality issue?

I already referred the posts here but this question is different. I don't wish to use categorical encoding. details given below I have a dataset of 3000 unique customers purchase data. The dataset ...
1 vote
2 answers
1k views

feature selection on training and test data

it is clear that feature selection (FS) have to be done separately on training and then on test data to avoid overly optimistic results. Lets assume that I have training set and test data set. Also ...
8 votes
0 answers
6k views

Is it ok to get negative Cosine Similarity using LSA? [closed]

I am getting negative cosine similarity value between two documents in Latent Semantic analysis. How should it be treated?
1 vote
0 answers
15 views

Help with terminology & methodology for a hierarchical (& imbalanced) classification problem

I have a dataset that I am not sure how to analyze, or at least I am not sure of the terms to read up on. I have 25 groups. Each group belongs to one of 3 locations. Each group consists of multiple ...
1 vote
0 answers
8 views

Temperature Lag calculation

I am working on a data science project on an industrial machine. This machine has two heating infrastructures. (fuel and electricity). It uses these two heatings at the same time, and I am trying ...
0 votes
0 answers
7 views

For a CNN Neural Network, why do we need to specify the number of nodes in the Conv2D function in Keras? [closed]

As I understand, in CNN, we are only doing dot product calculation on the image in the convolution layer. Below is an example of convolution code. ...
1 vote
0 answers
95 views

Calculating Shannon Information of Data Augmentation Strategies

I recently caught Andrew Ng's 2021 talk on MLOps (MLOps: From Model-centric to Data-centric AI). At 26:40, he talks about calculating the effectiveness of cleaning your data (training examples) vs. ...
143 votes
8 answers
148k views

What is the difference between off-policy and on-policy learning?

Artificial intelligence website defines off-policy and on-policy learning as follows: "An off-policy learner learns the value of the optimal policy independently of the agent's actions. Q-learning ...
2 votes
2 answers
388 views

Why does Hutchinson's trace estimator reduce computation complexity?

Given a matrix $A$, we want to compute its trace, in which we can use a trick name Hutchinson's trace estimator \begin{align} tr(A) = tr(A\mathbb{E}[\epsilon \epsilon^T])=\mathbb{E}[tr(A \epsilon \...
3 votes
2 answers
27 views

Neural networks - what is the point of having sigmoid activation function $\sigma(.)$ AND sigmoid g(t)?

Just so we're all on the same page, this is the classic neural network set up as I understand it: $$z_j=\sigma(\alpha_0j+\alpha_j^T x)$$ $$t=\beta_0+\Sigma^k_{j=1}\beta_j z_j)=\beta_0+\beta^Tx $$ $$ y=...
1 vote
0 answers
12 views

Using Variance of time series as input feature for time series clustering

I have a time series dataset, it is a data frame with 2000 rows and 1000 columns. Each rows is for one specific id and has a unique pattern. I want to clustering this data into multiple classes. Let ...
1 vote
0 answers
11 views

Hyper parameters to tune in Bayesian network?

In tree based models and neural networks, we can optimised the models by tuning the hyper parameters(such as: learning rate, number of neutrons.. etc). Is there a hyper parameters to tune in Bayesian ...
0 votes
1 answer
22 views

RFM Customer segmentation - Why Avg monetary value instead of total monetary value?

I am trying to segment our customers based on their purchase data. And I came to know about the RFM technique (Recency, Frequency and Monetary) through these posts here, here etc. Recency - How ...
0 votes
0 answers
13 views

Modeling demand distribution with selling constrained by stock

I'm working as a Data Scientist on a project where we are supposed to determine how many pieces of stock a certain retailer should have in each of its physical stores. The stock should be set on the ...
1 vote
1 answer
22 views

Can I convert a typical reinforcement learning problem to a supervised learning problem?

I'm not sure if I've understood correctly the whole point of reinforcement learning. In my point of view, the whole goal of RL is learning a policy that maps states to actions. Let us suppose that I ...
0 votes
1 answer
795 views

Test Accuracy same as Training Accuracy

I am building a prediction model using KNN. After experimenting the data using KFOLD Cross Validation technique, I've got the mean accuracy and applied them on the real model and it turns out that the ...
0 votes
0 answers
20 views
+50

Implementing kernel alignment for SVM algorithms

I am trying to understand and re-implement the results from Table 2 in the first Kernel-Target Alignment paper. The task that is being done is a simple classification task using an SVM with RBF ...
1 vote
2 answers
393 views

Literature on applying XGBoost to Time Series Data

I'm currently working on doing a time-series model with very limited data. However, most of the independent variables I have are not time-dependent, cross-sectional data. As such I want to apply some ...
12 votes
2 answers
2k views

Scaling the backward variable in HMM Baum-Welch

I am just trying to implement the scaled Baum-Welch algorithm and I have run into a problem where my backward variables, after scaling, are over the value of 1. Is this normal? After all, ...
-1 votes
2 answers
583 views

Mean Accuracy and Standard Error of the Accuracy for KNN Classification algorithm

The given below code snippet is from the assignment of online course IBM ML with Python. Here's the assignment. The used variable names :mean_acc and ...
128 votes
9 answers
69k views

Bias and variance in leave-one-out vs K-fold cross validation

How do different cross-validation methods compare in terms of model variance and bias? My question is partly motivated by this thread: Optimal number of folds in $K$-fold cross-validation: is leave-...
0 votes
0 answers
22 views

How does attention add expressive power to encoder-decoders?

I am learning about the attention mechanism for the first time, and the context in which it has been introduced (watching Lecture 8 of Stanford's CS224N) is that of language translation using seq2seq ...
0 votes
1 answer
239 views

With an LSTM, with training samples on 0->250, should it be able to extrapolate to unseen data(e.g. 250->500)?

I'm currently training on a simple dataset: Training: [0,1,2,3,4,5...250] Test: [251-500] My training cost and expected output decreases and seems correct. However, when I test the model, my network ...
0 votes
0 answers
11 views

Multiclass classification on top of given probability distribution from previous model

I have a multiclass classification problem that has multiple steps. Firstly, I am given a probability distribution over the classes by a base model for each sample. The task is to create a new model ...
1 vote
1 answer
56 views

Significance test for comparing different 10-fold cross-validated Machine Learning Regressions

Is there a recommended significance test for comparing different 10-fold cross validated regressions? For instance, I want to compare the performance of LASSO against Random Forest for my dataset. ...
0 votes
1 answer
29 views

how to generate data from beta-Liouville distribution?

I want to test my model following the beta liouville distribution, so as a synthetic data, I need generate data from this distribution. can anyone mathematically tell me how to calculate it? this is ...
3 votes
1 answer
188 views

Predictive Distribution in Gaussian Process Derivation

In Gaussian Process for Machine Learning (Rasmussen and Williams), pg 11, we are given the following predictive distribution: $$p\left(f_{*} | \mathbf{x}_{*}, X, \mathbf{y}\right)=\int p\left(f_{*} | ...
1 vote
1 answer
15 views

Neural networks - calculating output manually if $x_1=x_2=0$ . Should this be easy to do?

This is a problem question I'm trying to make sure I understand from a past paper (with no solutions). The R output is below. ...
0 votes
0 answers
6 views

Interpreting Shapley Values on Breast Cancer

I was analyzing Shapley Values on the Wisconsin breast cancer data set (binary classification). I applied it on Random Forest and on Ridge and Lasso Regression. However the summary plot seems to be ...
2 votes
1 answer
22 views

Does using grid search for hyperparemeters make test set redundant?

The purpose of train, validate and test data splits addresses the issue of data leakage when tuning for the model's hyperparameters. Does Grid Search then eliminates the need for test set? Because ...
2 votes
1 answer
23 views

1D cluster - Jenks optimization - Finding optimal number

I have a sample data variable shown below score 10, 11, 12, 90, 95, 97, 38, 37, 35 Instead of applying/binning data based on ...
0 votes
0 answers
11 views

Assessing importance of interactions between categorical features

The issues with using feature_importance of models such as XGBoost, or even using packages like SHAP or ELI5, is that the results are displayed in a way that doesn'...
2 votes
3 answers
19k views

How do I increase accuracy with Keras using LSTM

I will start with saying I am a complete beginner and doing this assignment for a class, and having some issues on how to get this to be accurate and (somewhat) show it's working! Can someone that ...
2 votes
1 answer
348 views

How to calculate explained variance of Self Organizing Map

Learning SOM recent days, but getting curious how does the explained variance of SOM is calculated. All the articles I have seen ignore this topic. Can anyone give some ideas?
-3 votes
0 answers
19 views

"Consensus" on Analyzing Mixed Continuous and Categorical Data in the field of Statistics? [closed]

I have been trying to determine the popular "consensus" as to how mixed continuous and categorical data (e.g. a dataset that has variables on income and gender) is generally analyzed in the ...
8 votes
1 answer
71 views

Feature engineering for sheet music

I have a large dataset of digitized music scores that I'd like to use as input to a network. Initially, I'm looking to train networks to identify key signatures, tempo, dynamics, etc. from the raw ...
2 votes
2 answers
1k views

Using one ml models output to choose another models input

I'm dealing with a low event rate problem (e.g. credit card fraud). I've balanced my data with SMOTE, and ran a neural net model (cross validated with recall as the measure). However my precision (...
-1 votes
0 answers
9 views

How can I train Mixture Density Neural Network? [closed]

I am learning Mixture Density Neural Network but it looks different from the usual neural network for regression problem. As far as I have understood from what I have read on the Internet, it gives ...
2 votes
1 answer
10 views

Is it possible to deal with datasets of graphs with different number of nodes in graph nural networks?

I'm dealing with a graph classification problem. In my dataset, each graph has som specific number of nodes. The number of nodes has a range of 1-1000 nodes. At inference time (after training), the ...
3 votes
1 answer
86 views
+500

How to force splits in decision tree to be distributed uniformelly in case of no dependency on feature?

I have targets ordered by a feature. I want to find a single split that minimizes a squared deviation (RMSE). For example, I have 100 values (targets) and it might be the case that, if I split them as ...
2 votes
1 answer
267 views

ML Method for directional forecast

I've uni-variate demand data (Weekly data for 2 years), and wish to do a directional forecast based on the data. Magnitude of the forecast is not important here, but directional accuracy is of ...
0 votes
0 answers
12 views

How do you approach a CNN problem?

I wanna create a machine learning model based on a region based CNN architecture (either RCNN, Fast RCNN or Faster RCNN). As an framework I wanna use Pytorch. I made a image containing apples and ...
2 votes
0 answers
56 views
+50

Are there smooth analytical penalty on leaves sizes for decision trees?

In a decision tree, when we search for an optimal split, we usually minimize root mean square deviation (RMSE). In addition to that we might forbid splits that give too small leaves (for example a ...
1 vote
0 answers
14 views

How to manage out of sample data in the long run?

For example, you are interested in testing an investment strategy and there is data from 1950 to 2022. So you split it into a train and test set, say 1950-2000 and 2000-2022. Then you build your model ...
1 vote
3 answers
2k views

What does Penalize a learning algorithm mean in Machine Learning?

I am new to Machine Learning and have taken Andrew Ng's course on Machine Learning. In one of the Logistic regression videos for binary classification for the error case where predicted value through ...

1
2 3 4 5
370