Questions tagged [modeling]
This tag describes the process of creating a statistical or machine learning model. Always add a more specific tag.
1,945
questions
0
votes
0answers
2 views
Output of randomForest and MeanDecreaseAccuracyValues
I have a question relating to the “randomForest” package in R. I am trying to build a model with ecological variables that best explain my species occupancy data for 41 sites in the field (which I ...
0
votes
0answers
8 views
Modeling question and the variable standardization
I'm seeking advise on the model variable standardization, supported by experience and the pros and con of variable standardization:
$$\frac{x_i-mean(x)}{std(x)}$$
There is a entry on this topic here ...
1
vote
0answers
8 views
Connection between Bayesian model selection and linear regression objective function
I'm currently studying machine learning using the book Introduction to Machine Learning (Alpaydin, 2014) and had a question. More specifically, this question is regarding a part of Chapter 4.8: Model ...
0
votes
0answers
14 views
Weightings using exponents
I've got this piece of work to do where we assign weightings to different variable to achieve a score:
var A = 60%
var B = 40%
var C = 20%
var D = 5%
The score was calculated as:
$(A^{0.6} x B^{...
8
votes
1answer
143 views
Recurring problem with retrospective data collection study designs I'm seeing
I've noticed a lot of medical research that I am involved in goes as follows:
Collect data on 300-1000 patients, including all sorts of baseline characteristics such as BMI, age, gender and then ...
1
vote
0answers
12 views
Autocorrelation at various aggregation levels
I have a time series sampled at a relatively high frequency that exhibits significant short and long-term memory effects.
Looking at the standard ACF, I can see highly significant negative ...
0
votes
0answers
27 views
If Maximum Likelihood estimation finds the best set of parameters for a regression model, then are likelihood ratio tests unnecessary? [closed]
My logic is this. Models are sufficiently defined by their parameters. Since MLE picks the n parameters that maximize the log-likelihood, then MLE picks the best model given n degrees of freedom of ...
2
votes
2answers
62 views
Is there a simpler probabilistic causal model that describes this data generating process?
Consider the following data generating process:
A person with gender male or female is selected from a population with probability $\alpha$ of selecting female.
The person is offered a drug to treat ...
-1
votes
0answers
27 views
How to choose the best regression model
For my statistics analysis, I try to understand the variables that influence collaboration of Europese universities with china. At this point in time, de DV = Chinese publications/total publications ...
0
votes
0answers
12 views
Exact inference in an approximate model as opposed to approximate inference in an exact model?
I remember hearing a while ago that it was more rigorous to perform approximate inference in an exact model as opposed to exact inference in an approximate model. I can’t now remember the reasoning ...
2
votes
1answer
54 views
Is setting a certain covariance structure between random effects and zeroing R equivalent to setting this structure exclusively in residual matrix?
I'm wondering whether setting, say, a compound symmetry covariance structure between random effects and setting the residual covariance to 0 is effectively the same as not using the random effects G ...
0
votes
0answers
23 views
Proper way to test proportions of data points output by a model
I have a dataset with a series $1000$ categoric observations. There are around 250 distinct categories, with $20$ distinct categories comprising around $55 \%$ of the dataset, thus most of them repeat ...
2
votes
1answer
35 views
Classical tests of hypothesis vs. regressions in small studies
I'm wondering if a study size is relatively small and for whatever reason you insist on doing inferential statistics, would classical tests of hypothesis be more appropriate than regression analyses? ...
1
vote
1answer
54 views
What is the outcome/coefficient of a LASSO? [closed]
A multiple logistic regression yield odds ratios, 95% CI's and p values which I understand. LASSO (logistic) seems to yield deviation and deviation ratios and no p values. I'm not sure how to ...
0
votes
1answer
24 views
What is the difference in between quasi-Poisson regression model and generalized additive models? On what basis do we choose these models? [closed]
I read some papers where some authors used one model to the other, I tried to find the underlying assumption behind the model but can't fully understand it. It is not clearly mentioned. One reason I ...
3
votes
1answer
60 views
Model selection with low N?
I've got a study which is kinda messed up by the design...Turns out I ended up with about 50 patients of which 80% have the outcome and 20% don't (binary outcome).
I've been turning in my bed for the ...
0
votes
1answer
26 views
Low N, unsure on what model to build [duplicate]
I've got a study with 55 patients having undergone surgery. 80% were happy with the surgery while 20% weren't.
I'm looking at predictors that may be able to predict surgery satisfaction.
The problem ...
0
votes
1answer
63 views
When should we avoid overfitting? [closed]
In the case of model fitting, if we are sure both training and test data have no noise, still must avoid overfitting? (training data may be insufficient)
0
votes
0answers
7 views
Interpreting AIC values [duplicate]
I have not much idea about statistics, but I know that one of the most used parameters to evaluate a model is the AIC value (the lowest the better according to what I read).
The thing is that in ...
0
votes
0answers
21 views
Non-stochastic vs Stochastic regressors and sampling distributions and causation?
I was wondering if I understand these correctly. Would an example of a stochastic regressor be weather? so when thinking about the sampling distribtuion and causality, I would think of repeated ...
2
votes
1answer
65 views
R formula for higher order polynomials and interactions, only allowing polynomial of degree 1 to interact
I am trying to build a (mixed) model using several predictor variables, and including some interactions and potentially higher degree polynomial versions of the continuous variables. The model formula ...
0
votes
0answers
26 views
Modelling distance distributions
Overview
I am building a model for a dependent variable which represents distance (from a more general perspective, my response can only take positive values). Moreover, it is imperative that I am ...
0
votes
0answers
32 views
Chi-Squared test for statistical analtsis
I'm comparing sonic logs against well logs at 0.5m increments. The soil is identified to be one of 6 categories: SOIL, FILL, GR, SA, SI or OM.
I am attempting to find whether the well data is ...
0
votes
1answer
10 views
Conv network gets worse when using more data
I'm trying to train a car image classifier based on a smaller version of AlexNet.
I'm trying to learn about conv nets and training models in general, and I've found something that, to me, seems ...
1
vote
1answer
35 views
How to deal with differently skewed biological data?
I have a single-cell data set with around 40 variables per cell (protein expression, all variables are measured simultaneously). The expression distributions for the single channels look quite ...
1
vote
1answer
43 views
Should I use the Diebold-Mariano test year-by-year or on the overall forecast?
I have built two models, one ARIMAX and one VAR, to compare against a baseline ARIMA model to predict a weekly economic time series of interest. I am primarily comparing the accuracy of my models ...
0
votes
1answer
31 views
Multiple regression with 2 correlated variables [duplicate]
I have a data set with a dichotomous outcome variable (surgery result good/bad) and two MR scan markers that are continuous (measurements on the scan).
Now if you have a large measurement in one ...
1
vote
0answers
34 views
Why ever use any of the tests of hypotheses when we can just regress everything?
So far in every data set I've worked with, regressing with only 1 predictor variable has yielded p-values extremely close to the one I get when running it with a test of hypothesis such as pearson's, ...
0
votes
0answers
11 views
Hypothesis tests via permutation importance on validation data
A model agnostic way to measure variable importance of a variable $x$ in a statistical model (originately proposed by Breiman) is as follows:
Select a scoring function $\mu$ to measure model ...
1
vote
0answers
22 views
Randomness in parameters per se captures the incomplete knowledge on the phenomenon: analysis in Bayesian models
I have been studying some books on uncertainty quantification for stochastic systems: Numerical Methods for Stochastic Computations: A Spectral Method Approach and Spectral Methods for Uncertainty ...
0
votes
0answers
6 views
technique/model to identify tracking issues?
I am learning Data Science and have decided the best way to learn is by application.
We have a tracking script on our website which fails to track 40% of visitors. Some of these could be due to ...
1
vote
1answer
11 views
Two level analysis, but mostly just one case per higher level - what to do?
There are 500 stores (n = 500), for which I want to model revenue.
Each store has data on their level, like size and age (the lower level) but also data on a higher level (let's say city level, ...
2
votes
1answer
37 views
Linear regression with hour of day variable
I have read related questions about linear regression models with hour of day variable, but they didnt help me at all, so I want to describe my doubt exactly:
I am facing a linear regression problem, ...
1
vote
1answer
26 views
How to model energy production capacity
I am doing a market research for wind energy production capacity (in megawatt) in an African country. I have data about the wind energy production capacity over the years. It looks like the following :...
3
votes
1answer
26 views
GLMM in seed germination study
I have an experimental design measuring germination of a single species of tree under different treatments. The treatments include; cattle grazing and no cattle grazing and rodents and no rodents. The ...
2
votes
2answers
52 views
What tests can I apply to following data in R?
We're doing our first project in R as our final grade, which is to find a dataset and test some hypotheses relating to the dataset. Our R knowledge is pretty basic and stuff we know are pretty much ...
1
vote
1answer
34 views
AIC for model comparison
I have three non-nested models with different predictors over the same outcome variable and I used AIC to compare their relative quality. However, I am very confused as to how to interpret the output ...
9
votes
2answers
823 views
Isn't it futile to try to fix multicollinearity?
Most of the advice on how to deal with multicolinear predictors tells you to eliminate them before fitting your model, using some criterium like VIF (Variance Inflation Factor). If I understand it ...
1
vote
1answer
29 views
How should I understand smoothing in functional data analysis from a modelling perspective (specifically for temperature data)?
The specifics of this question are that I am looking at daily maximum and minimum temperature from the GHCND data set obtained from NOAA's API and I view the temperatures observed at days in a year as ...
2
votes
1answer
39 views
Data analysis of unbalanced study - Interaction techniques in virtual reality
I'm planning a study on interaction techniques in virtual reality. That means I want to compare the performance of the participants on different interaction forms (e.g. selecting objects with a ray or ...
0
votes
0answers
13 views
Standardisation of predictor variable
I got a quick question that is confusing me.
A multiple regression analysis predicting job confidence from work demands, work support, and the interaction between them (WKDxSUPP) is to be conducted.
...
2
votes
1answer
40 views
Multiple observations from the same individuals - can linear regression still work?
Consider a project in which you have 120 participants answering a questionnaire about perceived quality and perceived price of several variations of a product.
Each participants gets introduced to a ...
0
votes
2answers
36 views
Model to be used with only median data and weights
I am finding a decent method to demonstrate if there is trend in the median age at diagnosis of multiple cancers. I do not have a breakdown of individual ages though, all I have is only (1) Median age ...
0
votes
0answers
10 views
Data Balancing for uplift model
When building an uplift model (Both using indirect methods such as the two models approach and direct methods such as class transformation and causal trees) how important it is to balance the data?
Is ...
3
votes
2answers
38 views
What is the difference between $ y = \alpha+\beta_1x_1+\beta_2x_2 $ and $ y = \alpha+\beta_1x_1+\beta_2x_2+\epsilon $ for a linear regression?
I am told that the equations in my question are different but don't really understand why they are different. One said the equation, $ y = \alpha + \beta_1x_1 + \beta_2x_2 $ is deterministic and the ...
0
votes
0answers
24 views
COPAR modeling in R
I was looking for a package supporting the modeling of a COPAR model by Brechmann et. all (see Reference below). However, I was not able to find one.
Using the packages VineCopula, Copula.Markov, ...
16
votes
3answers
946 views
Confused about Autoregressive AR(1) process
I create an autoregressive process "from scratch" and I set the stochastic part (noise) equal to 0. In R:
...
0
votes
1answer
29 views
Modeling Without Dependent Variable
I’m trying to figure out this problem where I want to calculate the probability of a set of people underpaying a service.
The service needs to be paid as a percentage of people's income. The issue ...
0
votes
0answers
8 views
What is threshold in ROC curve? [duplicate]
Whenever I read about ROC, people say that it is graphical representation of True Positive Rate value and False Positive Rate at various threshold.
Whenever I read in detail, people explain that ...
0
votes
1answer
11 views
What can I do when the values for my attributes is mostly 0
I was trying to do classification for census-income (KDD Data Set) (https://archive.ics.uci.edu/ml/datasets/Census-Income+%28KDD%29). The aim is to classify people who gain income >50 k and <50k. ...
