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
20 views

Variance of Guassian Products [duplicate]

Suppose I have to vectors $w$ and $x$, each of size $[512,1]$. Each element of $w$ and $x$ is an i.i.d sample from a Guassian Distibution with mean 0 and variance 1. So $x_i$ and $w_i$ follow $N(0,1)$...
0
votes
0answers
7 views

Question regarding feature correlation and feature selection for text classification [closed]

I am a complete newbie and I have a question regarding feature correlation and selection for a text classification problem. I am trying to classify a collection of text into a certain number of ...
0
votes
1answer
11 views

Size of training in Naive Bayes

I just started getting involved with Machine Learning and I decided to create a spam filter for my social app, using the Naive Bayes classifier. I'm following this guide: https://hackernoon.com/how-to-...
0
votes
0answers
19 views

Can a classifier A get better result than classifier B when learning from the output of B?

I had the following problem recently: I tried to reverse engineer a classifier $C_1$. $C_1$ is an unknown, already in production classifier which I can't access. I can only access the result on past ...
0
votes
0answers
9 views

How to calculate correlations of adjacent layers in a neural network

According to this and this blog post group convolutions as used in AlexNet, ResNext, etc. not only provide a more efficient and parallely trainable architecture but also enables the net to learn a ...
1
vote
1answer
38 views

KL divergence minimization

While reading Unsupervised Data Augmentation for Consistency Training, I came across an equation that describes the minimization of KL divergence. $$\min_\theta \mathscr{J}_{UDA}(\theta) = \mathbb{E}...
3
votes
0answers
97 views

A maximization problem, with motivation in machine learning

Consider the minimization problem described this paper. Let $f_{\lambda}$ be the minimizer. As a part of extending my work, I am able to show the following facts $$\lim_\limits{\lambda \to 0}\|f_{\...
0
votes
0answers
38 views

Benefits of ML in signal processing

There is plenty of research on ML in signal processing. The majority of it, so it seams to me, is about showing feasibility of ML-based receivers (end-to-end or individual functional blocks of). To me,...
1
vote
0answers
14 views

Should initial features be added as input of meta-learner in stacking?

I'd like to know your opinion and reference on adding initial features (i.e. the ones used to train the weak learners) to the input of the meta learner (aside of the predictions from weak learners) ...
0
votes
1answer
20 views

Range of Values for Random Forest Mean Decrease in Accuracy

When calculating variable importance using the unscaled (Scale= FALSE) permutation variable importance, what is the range of values you can get? Is it expressed as a percent (e.g. a value of 0.1 ...
0
votes
0answers
20 views

How to perform machine learning on data with many trials on individual subjects

I have an EEG dataset. The data are epoched around an event of interest and are of the form channels x timepoints x trials. There are many of these types of datasets for several subjects. Because ...
1
vote
0answers
12 views

Supervised learning in non-stationary environments?

For real applications, concept drifts often exist, i.e., the relationship between the input and output changes overtime. I'm wondering what are the most common methods to enable neural networks to ...
1
vote
0answers
16 views

Online Optimization - Regret in Absolute Error

In the online convex optimization literature static regret is defined as $\sum_{t=1}^{T}\left(f_t\left(x_t\right)-f_t\left(x^*\right)\right)$ where $x^*=\arg min_{x\in\mathcal{X}}\sum_{t=1}^{T}f_t\...
0
votes
0answers
21 views

What is the difference between validation and cross validation?

I hear the dataset is divided into three parts 1-training 2-validation 3-test in the 2- validation Do you mean cross validation or what difference does it make?
0
votes
0answers
3 views

Evalution Metric for Recommender with one Relevant Document

Suppose I have a bunch of user session data. For each user session, 5 rows are created. Each row contains the user id, item id and whether or not they selected that item. For example : ...
0
votes
0answers
13 views

Robust Expectation-Maximization?

The Expectation-Maximization (EM) algorithm is useful for applying the Maximum Likelihood Estimation (MLE) when there exist latent (hidden) variables in the model. However, when dealing with outliers, ...
0
votes
0answers
8 views

Order of continuity of an ANN approximation dependent on the activation functions used?

If I have understood this correctly, a result from Hornik et al.'s Universal Approximation of an Unknown Mapping and Its Derivatives Using Multilayer Feedforward Networks essentially states that, if ...
1
vote
0answers
15 views

Stop-gradient operator in vector-quantized variational autoencoder

The objective function in VQ-VAE (Eq. (3) here) contains $$\left\lVert \mathrm{sg}[z_e(x)] - e \right\rVert^2 + \left\lVert z_e(x) - \mathrm{sg}[e] \right\rVert^2,$$ where $\mathrm{sg}$ is the stop-...
0
votes
1answer
31 views

have new outliers after capping

I'm trying to cap outliers in a column my pandas DataFrame. Here's the boxplot for a column of my original data. boxplot for a column of my original data So, using code from this stackoverflow answer, ...
1
vote
1answer
24 views

What do P and Q refer to in the Minkowski distance?

Wiki gives some explanation and a figure about the Minkowski distance: The Minkowski distance is a metric in a normed vector space which can be considered as a generalization of both the ...
0
votes
0answers
25 views

Can feature importance be measured in machine learning models by selectively removing predictor variables?

In addition to techniques like measuring permutation importance, could feature importance be estimated by sequentially removing predictor variables from the training data, rerunning the model, and ...
0
votes
0answers
16 views

What do you think is the best method for prediction for network structures

Open Ended Question Here. Suppose that you have 10000 samples for 100 binary explanatory variables. Approximately every explanatory variable has a fixed probability of being 1 (say 25%). The outcome ...
0
votes
2answers
18 views

Working mechanism of discriminator in text to image synthesis GAN

I have the following architecture of discriminator in text to image synthesis where the image is convolved to lower dimension and concatenated with the text . My question is what is the use of ...
0
votes
0answers
15 views

Why my Deming Regression line change so much when switching variables? If they seem to be a linear relationship betwen them?

I am trying to fit a line that best predicts the production of energy Y given the speed of wind X, a typical Y = xm + b , using deming regression. I am looking for the slope and the intercept of that ...
0
votes
1answer
30 views

Manually adjusting forecasting model bias

I am trying to build an efficient forecasting model to predict sales in the future. I managed to obtain a first pretty solid model using a LSTM network. However, it wasn't sensible enough to large ...
1
vote
1answer
44 views

What is the common ingredients of a machine learning algorithm?

considered KNN, it seems that a machine learning algorithm does not have to have weights, a training process, loss function or optimization, so, what is the common ingredients of a machine learning ...
1
vote
1answer
54 views

Differences between A Statistician and a Data Analyst In Industry

What is the Difference in the job of a Statistician and a Data Analyst In Industry? My take is that although both analyse data, a Statistician deals with the more theoretical aspects of data such as ...
8
votes
2answers
365 views

Does KNN have a loss function?

I didn't find a definition of loss function on wiki in the context of machine learning. this one is less formal though, it is clear enough. At its core, a loss function is incredibly simple: it’s a ...
1
vote
1answer
44 views

Why does Perceptron use L1 norm as its error function?

Usually, people use L2 norm as a machine learning error function. per wiki the error function for a Perceptron model is ${\displaystyle {\frac {1}{s}}\sum _{j=1}^{s}|d_{j}-y_{j}(t)|}$ why is that?
-2
votes
0answers
13 views

Normalizing data in classification [duplicate]

i am following a classification tutorial and i encountered this line of code and i dont understand it . it says that we are going to normalize our data. anyone with knowledge please explain ...
0
votes
1answer
15 views

Relation between size of parameters and complexity of model with overfitting

I'm reading the book Pattern Recognition and Machine Learning by Bishop, specifically the intro where he covers polynomial regression model. In short, let's say we generate $10$ data points using the ...
1
vote
0answers
36 views

To what extent will results change when fitting random forest structurally vs in batch?

I am facing a problem where I am not sure how to apply the random forest algorithm. Suppose my target variable is $Y$ and I have some $X_1, X_2, \ldots, X_p$ as predictors. I am applying the random ...
0
votes
0answers
14 views

Unsupervised Learning vs. Supervised Learning in Natural Language Processing

I don’t quite understand the difference between supervised vs. unsupervised learning in Natural Language Processing, when trying to predict the next sequence of word in a sentence. Could someone ...
0
votes
1answer
29 views

Get individual features importance with XGBoost [closed]

I have trained an XGBoost binary classifier and I would like to extract features importance for each observation I give to the model (I already have global features importance). More specifically, I ...
2
votes
1answer
34 views

How to combine different results of a stochastic classifier

I'm writing a paper about a machine learning-based system and using CNNs on a GPU cluster to compare two methods of feature engineering. Due to the non-deterministic nature of the algorithm, I couldn'...
0
votes
1answer
27 views

Different terminologies for ground truth

It's general theory that a supervised learning approach can only get as good an accuracy as the quality of the ground truth. The ground truth is considered the best possible annotation. However, ...
0
votes
0answers
12 views

Why we really need the concept of “Local” Rademacher complexity?

Recently, I have been studying High-Dimensional Statistics: A Non-Asymptotic Viewpoint written by Martin J. Wainwright. In this book, the author uses a special complexity measure which is called Local ...
0
votes
1answer
61 views

Predict the price of single part after running the ML algorithm

I am working on a problem to predict the price of a mechanical part. I collected all the necessary data and variables.($1200 \times 25$) I have build multiple models using $log$ transformations of ...
1
vote
1answer
46 views

Specifying separable covariance functions for 2D gaussian process regression

I would like to fit a gaussian process regression with two input variables. But I am not sure how to construct or interpret the covariance function with multiple input dimensions. There are different ...
0
votes
0answers
17 views

Interpretation of F-statistic and pvalue Anova [python]

Dataset I tried both methods of finding f-statistic and pvalues of my categorical features. I sort the table in highest f-statistic values 1st and not the pvalues. Is this the correct way to find ...
6
votes
2answers
139 views

Why do lots of people want to transform skewed data into normal distributed data for machine learning applications?

For image and tabular data, lots of people transform the skewed data into normally distributed data during preprocessing. What does the normal distribution mean in machine learning? Is it an ...
0
votes
0answers
10 views

Approximate Value / Policy Iteration (Reinforcement Learning)

I am reading Markov Decision Processes in AI : about Approximate Dynamic Programming. Would you like to explain the rationale for introducing the API algorithm, how it compares to AVI ? How would ...
0
votes
0answers
21 views

Probability vs recall for time series classification task

I am working on the time series classification task that focuses on predicting a fault. I framed the problem as a multi-step forecasting problem, where my goal is to predict to the class at ...
2
votes
1answer
26 views

Optimizing parameters of fully connected layers

I have a conceptual problem choosing the best strategy for training a neural network. So, let me explain my situation. I have trained an autoencoder on a huge (unlabeled) dataset using train and ...
1
vote
1answer
62 views

Two dumb questions about standardization and overfitting [closed]

The following two questions may seem to be dumb, but I could not figure out reasons to convince myself. Question 1 Why neural networks (or more generally, any machine learning models) tend to ...
0
votes
0answers
38 views

How can I make input data for logistic regression?

I have dataset with 6 category variables. each category variable shows some machine error codes ranging between 0 and some big number like 50000. I want to try Logistic Regression to build a ...
1
vote
2answers
45 views

Compare how fast two time-series grow

I collected data from Pubmed containing articles published about two different topics. The first is a yearly series of articles about omega 3 supplements. The second is a yearly series of articles ...
0
votes
0answers
12 views

Big difference in random forest test subset error and other subsets

I'm using R "randomForest" package for predicting stock prices. I have more than 3000 observations and 90 columns. I have excluded my last 150 days from data set and divide the rest of my data to ...
1
vote
2answers
48 views

How to get the false positive? “1 - true negative” or “1 - true positive”?

wiki gives this Drug testing Example to illustrate Bayes' theorem ${\displaystyle {\begin{aligned}P({\text{User}}\mid {\text{+}})&={\frac {P({\text{+}}\mid {\text{User}})P({\text{User}})}{P(+)}}\\...