Stack Exchange Network

Stack Exchange network consists of 175 Q&A communities including Stack Overflow, the largest, most trusted online community for developers to learn, share their knowledge, and build their careers.

Visit Stack Exchange

Questions tagged [metric]

A metric is a function that outputs a distance between 2 elements of a set & meets certain strict criteria (some 'distance' functions are not metrics).

0
votes
1answer
36 views

Why is Pearson correlation is not an effective metric?

I found this statement in some documentation but I could not make sense of it. "Correlation is not a good metric for regression because it is scale and offset invariant". I understand that ...
0
votes
0answers
12 views

Set Similarity Measure that accounts for size

Suppose I have two sets, A and B where 10|A| = |B| and A is a subset of B. For the case where |A| = 1 the Jaccard similarity of the sets will be 0.1. For the case where |A| = 100 the Jaccard ...
0
votes
0answers
3 views

Comparing which histogram has overall low cost

Let's say there are two histograms which basically is constructed from array of numbers which is measured by, repeatedly performing a task by two different methods and individual numbers are time ...
0
votes
0answers
30 views

Is there a name for “normalized accuracy” as a statistic?

In short: I'm using a statistic representing the "normalized accuracy" of a confusion matrix. Is there a formal term for this? $$\text{normacc} = \frac{\text{acc}-\text{thacc}}{1-\text{thacc}}$$ ...
0
votes
0answers
19 views

Does Sørensen–Dice Coefficient (Dice Score) only account for true positives?

I'm working in a project on medical image segmentation which uses the Dice Score as part of the loss function, but I got some doubts with the commonly adopted implementation. The definition of Dice ...
1
vote
1answer
36 views

Difference between Euclidean ,Pearson, Geodesic and Mahalanobis distance metrics

Given a set of samples $X$. We are tasked to find an appropriate distance metric for $X$ from the given options which are Euclidean Pearson Geodesic and Mahalanobis distance metrics. To solve this, ...
1
vote
1answer
23 views

Clustering data with covariance for each point

I am looking to cluster data points that each have a covariance around itself (based on some function of its neighbourhood, but how I got it is not important). I would like to use the covariance to ...
0
votes
1answer
30 views

ROC-AUC score in sklearn

I'm trying to understand ROC-curve and AUC characteristic for it and found that behaviour of sklearn function roc_auc_score ...
0
votes
0answers
13 views

Is there a difference between Hausdorff Distance and Discrete Frechet Distance when working on time series?

I'm currently doing a little research on which kind of distance metric is the best for comparing the time-series I'm working with. To be clear, I'm doing this for a computer science internship, and I ...
0
votes
0answers
37 views

Why am I getting r2_score on test set as negative?

I wanted to initally test out without dropping any features (Redundant features such as ID are dropped). data_source -> https://archive.ics.uci.edu/ml/datasets/Automobile This was my procedure: 1) ...
0
votes
1answer
19 views

Creating a custom metric based on diverse feature scalse

I would like to create a custom metric that takes into account multiple features and produces a certain score. However, the issue is that I have data that has many different data types, such as ...
0
votes
0answers
18 views

How to interpret/compare R2 score(s)?

I know that an R² score of 1 is a perfect fit of the model to the truth, a 0 is an constant output regardless of the input, and that negative values are possible when the output varies, but there is ...
2
votes
0answers
26 views

Distance / Metric between two regression models? [closed]

I wonder if there is any theory or work about the "similarity" of two regression models. For example, if it is linear regression, the "similarity" could be defined by the l-2 distance between the ...
0
votes
0answers
18 views

Multiclass classifier with undefined prediction, how to calculate metrices

I build a multiclass classifier. I want the classifier to predict a few samples with little false positives, rather then many samples with lots of false positives. Therefore I want to choose a ...
0
votes
0answers
19 views

Metrics for comparing a multi-class model vs. a multi-label model?

The dataset is a multi-label dataset, where each item can have more than one labels. I first trained a multi-class classifier by randomly select the label for each item at each iteration, and ...
0
votes
0answers
38 views

Computing and estimating the EER on an entire dataset

I have reproduced "Generalized End-To-End Loss For Speaker Verification". It describes a method to create a deep learning model that can derive an embedding (a vector of 256 float values) that ...
1
vote
0answers
60 views

How do you / can you compare Bayesian vs. frequentist regression models?

I am working on a regression model to predict a target variable in a dataset with over 100 features. Three different regression models are defined and fit in order to compare their performance using $...
0
votes
0answers
21 views

Replacement for angular distance metric

I am looking for a distance metric that could be used instead of cosine/angular distance for high dimensional data. Metric that is limited the same way as cosine/angular distance is would be great. ...
0
votes
0answers
19 views

Consistency metric explanation

I am trying to understand a bit more about the consistency metric (to understand how consistency-based subset evaluation works). I find on this paper the following equation : $$\text{Consistency}_s =...
1
vote
1answer
124 views

Suitable performance metric for an unbalanced multi-class classification problem?

I have an unbalanced multi-class classification problem with the following class distributions: Class 0: 17.1% Class 1: 63.2% Class 2: 19.7% I am using scikit-...
0
votes
0answers
18 views

Mahalanobis: Covariance or correlation? [duplicate]

Is the Mahalanobis distance using correlations or covariances (among two vectors) to determine the similarity? I know that both are quite the same and differ only by a division of a standard deviation ...
0
votes
0answers
10 views

Which performance metric to use for stratified data? [duplicate]

I'm trying to classify a data into 3 classes (supervised), one of which is heavily underrepresented in the data set. In order to combat this imbalance, I decided to stratify the data. Now I want to ...
1
vote
1answer
26 views

Is recall a relatively meaningless metric in a balanced dataset?

Just looking for a sanity check here. Leaving aside precision, is talking about the recall of a binary classification algorithm sensible where 50% of the cases presented to it are positive and 50% ...
0
votes
0answers
10 views

Machine learning: eval_function for multiclass classifcation

I have checked What are the measure for accuracy of multilabel data? but I think the discussion somehow confused between loss and evaluation function. Say I have a multi-class classification problem ...
1
vote
2answers
48 views

Metrics for unbalanced classes [duplicate]

I have been looking for good metrics on this data set I am working, however it is highly unbalanced. It has a total of 8 categorical classes, one of them is responsible for ~40% of data, another for ...
2
votes
2answers
36 views

Localized distance function on sequential binary data

I am trying to find a good distance function for sequential data that is all binary. For now, I am using Edit distance however I have some more domain-specific knowledge that I would like to ...
0
votes
0answers
20 views

Quality metric of sampled time series data

I have a time series that has too many points. I sample one in every 100 points, in order to reduce the amount of data I need to transmit from my measurement device. What accuracy metric can I use ...
1
vote
1answer
135 views

Imbalanced Test Data

I have an imbalanced (1:5) training and test set with only two classes and have oversampled the training set with SMOTE so that the class ratio is 1:1. The ML model gives values over 0.7 for accuracy, ...
0
votes
0answers
22 views

Measure for evaluating a generated ranking of items, when best k items are known

I am evaluating different kinds of feature selection algorithms. Some return a ranking of variables based on their usefulness (to predict a target variable, etc.). Lets say there are n variables in ...
2
votes
2answers
817 views

Loss function and evaluation metric

When building a learning algorithm we are looking to maximize a given evaluation metric (say accuracy), but the algorithm will try to optimize a different loss function during learning (say MSE/...
0
votes
0answers
91 views

Best performance metric for highly imbalanced dataset f1 score vs kappa vs AUROC

I have highly imbalanced data (like fraud detection). I usually use f1 score to evaluate model performance. But I also saw people recommend AUROC and cohen's kappa. I'm seeking expert opinion on what ...
0
votes
0answers
16 views

Correct way of calculating cohort-based daily retention metric

Background: I need to calculate retention of users observed in a system. I will use this metric to estimate the Bayesian probability of said system's retention levels in the future. This question is ...
0
votes
0answers
7 views

Does zero-padding truth and estimate deteriorate the NRMSE/NRMSD?

I feel like I am missing an important point, but maybe the NMRSE (or NMRSD) is just counter-intuitive. This is the equation that I use: $\textrm{NMRSE} = \frac{\sqrt{\sum_i \left( \textrm{est}_i - \...
1
vote
1answer
557 views

different trends in loss and AUC ROC metric

I am training a deep neural network for a binary classification I am using binary_crossentropy as loss and area under the roc curve as performance metric as ...
0
votes
0answers
20 views

is there a measure for the roughness of a contour plot

There has to be a measure for the difference between "instantaneous" change of "energy" along a line in a space compared to averaged changed of energy along a line. I could take a smooth surface in ...
1
vote
1answer
36 views

Can two imperfect rankings be combined to produce a better ranking?

So I've got a dataset that can be ranked in two noisy ways $R_1$ and $R_2$. Let's call $R_1$ and $R_2$ functions from the dataset $x \in R^N$ to a real number between 0 and 1. Since they are rankings ...
0
votes
1answer
183 views

Assess the dependence of LDA on the random seed

New to latent Dirichlet Allocation (LDA), I would like to be sure that my output (in the first the step, the word-per-topic probabilities) depends on the input merely, and is (somewhat) stable ...
2
votes
0answers
55 views

Can you compute errors bars from Precision and Recall?

I am performing an object detection task for counting cars in an image. I have the confusion matrix (TP, FP, FP, TN) of the model. I guess TN is just zero in this case, as we aren't detecting where a ...
1
vote
0answers
17 views

How correctly overcome problem with an infinity MAPE? [duplicate]

Should I choose an other metric or is the way to handle this problem?
2
votes
0answers
26 views

How are mean results on benchmark obtained when training neural networks?

In most neural network papers, networks are trained on a known database where state-of-the-art performance is known ("benchmark"). Whatever metric is chosen to illustrate the network's performance, ...
0
votes
0answers
50 views

clustering: what metrics can be used to measure both intra-class variance and inter-class variance, and what are their complexity?

I was considering using the silhouette score, however its complexity is $O(mn^2)$ which can quickly become prohibitive with the size of the dataset.
0
votes
0answers
20 views

Lorenz curve and inequality between sports teams

I was wondering if you could develop a metric to study the inequality for a given stat such as how unequal points are distributed throughout a team, or how unequal rebounds are distributed throughout ...
2
votes
1answer
196 views

Validation set early stopping on custom metric

I am wondering whether it is ok to monitor validation set performance using a metric which is not optimized by the training algorithm, but which makes more sense in your domain. As a concrete example,...
1
vote
1answer
36 views

Can Negative Predictive Value and Positive Predictive Values be the same?

Is there any scenario where a negative predictive value and a positive predictive value would be the same? Specifically, when using a neural network for binary prediction. Can this be a sign of ...
3
votes
1answer
170 views

Does it make sense to use an Early Stopping Metric like “mae” instaed of “val_loss” for regression problems?

I am performing a regression on a Dataset and try to replace a mathematical Model with a Neural Network. To avoid overfitting I decided to use the Early Stopping Callback Function of Keras. So far I ...
0
votes
0answers
20 views

Sentiment analysis on IMDB database and choose of metric [duplicate]

I am currently working on an extract of the IMDB available on http://ai.stanford.edu/~amaas/data/sentiment/ . Since i try to predict the label (between 1 and 10) related with each review i face a ...
1
vote
3answers
466 views

When is an AUC score misleadingly high?

I have an algorithm which gives an AUC (area under the receiver operating curve) of 0.94. I mean, this is amazing, but... probably too amazing, considering the difficulty of the task I am working on. ...
0
votes
0answers
221 views

Jaccard distance for sets of different sizes

Is the Jaccard distance still a metric if applied to sets of different sizes? Or does it require the input sets are of the same length?
0
votes
0answers
27 views

Can cross entropy be employed for measuring effective mapping of cross-modal data?

I am working on a new metric for cross-modal retrieval which can measure the effectiveness of mapping two modalities on a manifold. However, the usual approach is to employ a distance metric and not ...
0
votes
1answer
43 views

the score to hope for when evaluating model by MAE, MSE or RMSE

when doing evaluation and optimization of model by MAE, MSE or RMSE what should we look at and compare our score to, as a baseline or acceptable score for our model. should we look at the best ...