# Questions tagged [predictive-models]

Predictive models are statistical models whose primary purpose is to predict other observations of a system optimally, as opposed to models whose purpose is to test a particular hypothesis or explain a phenomenon mechanistically. As such, predictive models place less emphasis on interpretability and more emphasis on performance.

3,188 questions
Filter by
Sorted by
Tagged with
10 views

### understanding fisheries model output tag and release

I am reviewing (for myself) a publication that has analysed data from an experiment to see what effect underwater seismic has on catchability of lobsters. de Lestang et al (2024) Fisheries Research ...
100 views

### How to generate 95% prediction interval around predictions from ML model?

I have predictions from an ML model and would like to generate 95% prediction intervals around each prediction generated from the model such that I can claim that these are the plausible range of ...
• 1,037
27 views

### Very Specific Plateaus in Time Series Data

I am looking at time series data of the depth of water in different pipes. There is a rare occurrence where extreme amounts of water are trying to get into the pipe, but since it is full the water ...
36 views

### Forecasting Survival Analysis

I use the Kaplan-Meier estimator to represent survival functions between two groups. Suppose I have X events at a given time t. How can I predict time t+k to obtain X+i events? As with time series, is ...
28 views

### SHAP values under multicolinearity/feature dependence

My task is to explain individual predictions, but having read the original paper and sifted through the internet, I am still unsure whether using something like TreeSHAP can help me with the situation ...
• 21
10 views

### Can one determine the number of forecast/prediction steps in a VAR on a priori grounds?

Context of my question: I am running a vector autoregression (VAR) model using two time-series of equal length (n ~ 750 data points). The lag was chosen based on the Bayes information criterion (BIC) ...
• 307
327 views

### What is the best epoch to evaluate the test images?

I created a training, a validation and a test set for an image classification task. Then, I did training using the training and did evaluation on validation set. So, the next step is to evaluate the ...
• 53
24 views

### Efficient prediction using Lightgbm/XGBoost when varying single feature keeping the remaining constant

Assume we have a pre-trained Lightgbm/XGBoost model $f$ dependent on the feature matrix: $$X=\left[z, C\right]$$ Here $z$ is a single feature column and $C$ is the remaining feature columns. I need to ...
• 293
1 vote
36 views

### Surival Prediction - Train/Test data vs Production data

I have a need to create a churn prediction model and it seems like a survival model fits the bill since my data is right-censored (there are many customers who have yet to churn, or in other words, ...
20 views

### Using a model to evaluate over or under-priced rental prices for the same apartments used in training

If I have a machine learning model which predicts the rental prices of apartments, can I use the model once complete to analyse the prediction for the same apartments I used to train the model so I ...
• 83
1 vote
37 views

### Predictive Maintenance of factory parts

I'm training a model to perform predictive maintenance of a particular part in a factory. I have performed data cleaning like removing the null, duplicate values, removing the highly correlated ...
• 13
13 views

### A causal path problem of prediction models

I am doing a study on predicting the 10-year risk of diabetes in people without diabetes using factors such as glucose, HbA1c, and triglycerides. The blood sampling time is not limited to fasting. In ...
• 47
20 views

### Fixed-effects using demeaned data: Why are the demeaned predictions different from original data w/ fixed effects?

I'm hoping someone can help me understand why a fit regression model making predictions on panel data with group fixed-effects, outputs different results than the same model predicting on the same ...
• 123
21 views

### Is Mean Square Prediction Error acceptable to use if predicted values are continuous but actual observed values are discrete?

I would like to compare the predictive power of 2 models. The models are meant to model count data, so the actual observed values are discrete. However both models are designed such that they output ...
• 133
32 views

### Neutral number in XGBoost algorithm prediction

Does a concept of "neutral number" in machine learning algorithms exist? To make it clearer: suppose we have a logistic regression with only one feature, the "neutral number" is ...
16 views

### Standard error of RMSE and differences in RMSE

I have a set of models $M = \{1, ..., m, ..., K\}$, and for each I am calculating RMSE on out-of-sample data as standard: \mathrm{RMSE_{m}} = \sqrt{\frac{1}{N} \sum_{i=1}^{N} (\...
• 185
1 vote
31 views

### Evaluating estimator of expected value plus variation

I know that for a typical, the estimator can be evaluated based on the mean squared error (MSE) of the predictions. How can I evaluate an estimator that instead gives a value that is the prediction ...
21 views

### Which is the correct regression model for predicting the association of climate with Julian days nested within decades?

Below is a reproducible example: ...
• 23
486 views

### What distribution should I use to predict three possible outcomes

I am 70, left school at 14 but took to maths a few years back to ward off dementia so please excuse the naivety of my question. I have been using Poisson distribution to solve my problem but I dont ...
• 151
1 vote
42 views

### Optimal performance measures for species distribution models with presence and pseudo-absence data

I am currently building species distribution models (SDMs) (or ecological niches) using machine learning algorithms to predict the potential spatial distributions of animal species based on ...
23 views

### Are diagnostic tests and clinical (risk) prediction models in medicine essentially the same thing?

In medicine, are diagnostic tests (e.g. covid test, HIV test, ...) and risk prediction models essentially the same thing? If not, in which aspects do they differ (from statistical point of view)? ...
• 3,989
1 vote
50 views

### Scaling Out-of-Sample Forecasts in a Model with Normalized Variables: Reverting to Original Scale

I'm working on making forecasts using a model where variables were scaled by $$\tilde x_i = \frac{{x_i - \text{mean}(x_i)}}{{\text{sd}(x_i)}} ,$$ and I've saved the mean and standard deviation. Now,...
• 301
135 views

### Using Predictive Value Confidence Intervals to "Predict" Outcomes

Here's the quick version: Say I have a confusion matrix with the following data based on a proficiency cut score on a pretest and outcomes on (passing/failing) a class. The cut score was determined ...
34 views

### Is there any test I can apply to the data to tell whether the adaptive LASSO or the LASSO is likely to perform better in prediction?

Is there a. test I can perform on a sample that will tell me if coefficients estimated using the LASSO, the adaptive LASSO, or the relaxed adaptive LASSO are likely to give better (in the mean squared ...
• 3,157
1 vote
26 views

### Multiple Imputation for Missing Outcome Data

I have spent an extensive amount of time trying to understand the possible role of MICE in helping to "fill in" missing outcome data. I am relatively new to both multiple imputation and ...
• 11
16 views

### Logit model not predicting any values < 0 despite class imbalance

I am building a logistic regression model to identify potential channelling factors that predict whether a patient will initiate of one of two antidiabetic drug classes at a specific stage in their ...
• 51
184 views

### Back Transforming log-log Model for Prediction

I have a model that is log-log and I would like to make raw predictions of $Y$ with it: $\ln(Y) = B_0 + B_1\ln(X)$ All answers and articles I have found concerning back transforming for prediction ...
118 views

### Choosing Between Intercept-Only and AR-NN Models: Justified to not use the model with the lowest RMSE/MAE?

I have created two autoregressive models for forecasting: a basic intercept-only model and an AR-NN (autoregressive neural network) model. Both models show similar performance based on recursive one-...
• 301
57 views

### How to Predict a Growing Time Series with Changing Slopes in Python?

I have a univariate time series dataset that represents a continuously growing trend, but the slope changes at different intervals. I want to predict future values of this time series. Here are some ...
46 views

I wish you good health and the best in life (whatever that means for you in particular). Context I want to model aggregated (monthly) tornado counts in the United States for a question on Metaculus. ...
31 views

### Sufficient number of data to determine whether a model fits the data well

I was wondering if, chosen a proper mathematical model, there is a minimum number of data that allows us to state if the model fits well the data or it doesn't. I'll explain better my question. Let's ...
1 vote
41 views

### Time Series Prediction with Hidden Variables [closed]

I have time series data sampled every 10 seconds. I want to predict a target variable 𝑦 for 5 steps ahead using the input variable at the current time along with 10 past lags. My problem is that ...
• 149
1 vote
54 views

### How to use a multi-linear regression to forecast meaningful values

I have built a multi-linear regression model based two predictors $P_1$ and $P_2$ to predict $Q$: $$q = A + Bx_1 + Cx_2 + Dx_1^2 + Ex_2^2 + Fx_1*x_2$$ where $x_1$ and $x_2$ are the ...
157 views

### Estimate the probability that a crack length exceeds a threshold value after N cycles

The length $a$ of a crack after $N$ fatigue cycles $N$ is $$a(N;C,m) = \left(a_0^{\left(1-\frac m 2\right)}+ C\left(1-\frac m 2\right)B^mN\right)^{\frac{2}{2-m}}$$ where $a_0$ is the initial crack ...
• 18.2k
1 vote
171 views

### Determining an optimal level of aggregation that balances accuracy and granularity

I am looking for ideas for aggregating prediction outcomes in a way that maximizes the number of classes while minimizing classification error. As a motivating example, say I'm working on a prediction ...
22 views

### Can I restrict the predicted values in Partial Least Squares model?

I'm using the R package pls, and I want to use the Partial Least Squares method to create a prediction of my data. The input I'm ...
22 views

### should the standardisation of numerical variables be carried out before or after the rebalancing technique of the target variable?

I am dealing with a classification task of a binary target variable (company failure prediction yes or no) for a university project. I was wondering, should the standardisation of numerical variables ...
15 views

### How to improve prediction quality of LSTM model

I am trying to train an LSTM-based model in MATLAB to predict 365 next values given 365 previous values of a time series. Input shape is (1000, 365) and output shape is (1000, 365) i.e. there are 1000 ...
21 views

### Valid covariate definitions that respect temporal relationships in regression models

I am trying to gain a better understanding of reasonable assumptions to make and valid ways to define covariates in regression models. I am putting aside survival methods where time is fundamental to ...
• 771
1 vote
57 views

### An error occurred when using the xgboost as a classifier for hiclass [closed]

Bellow it's my example when using the xgboost classifier for hiclass. My question is specifically directed to the hiClass Python package for hierarchical classification. I would like to model the ...
• 21
165 views

### Brockwell/Davis seem to say more persistence implies better predictability---do I have a counterexample?

Brockwell/Davis, Introduction to Time Series and Forecasting, p. 40, write (notation slightly adapted; please refer to screenshot below) The best linear predictor $l(Y_{T})=aY_{T}+b$ for a stationary ...
• 34.2k
1 vote
52 views

### Which statistical model is suitable?

I have the results of a survey of $n=132$ patients with their socio-economic profile and their spending behavior on mobility-coins (my thesis topic). In the survey, we asked people how they would ...
• 11
121 views

### How should I be interpreting martingale residuals and zph?

I'm new to survival analysis and am practicing on the AIDS Clinical Trials Group Study 175 Data from the UCI Machine Learning Repository. After using R to fit the Cox PH model, I made a plot to look ...
• 41
108 views

### Variable selection strategies

I'm wanting to learn more about current best-practice approaches to variable selection, in the context of the reason for creating the model. I don't have much experience with modern methods like lasso,...
• 771
129 views

### Is it possible to evaluate causal algorithms on real world observational data?

Lot of times I get asked to use causal algorithms (e.g. algorithms estimating intervention results, or in general causal inference algorithms) and to compare them against non-causal prediction ...
• 460
97 views

### Are stationary processes non-predictable, and non-stationary ones predictable?

I am reading A canonical analysis of multiple time series by Box and Tiao (1977). In the abstract of the paper, the authors mention: The least predictable components are often nearly white noise ...
• 489
24 views

### Building an algorithm to use Bayes conditional probability in conjunction with association rules to build predictive models?

Association rules are a data mining technique used to identify interesting relationships between variables in large datasets. These rules take the form "if-then" statements, where the ...
• 11
37 views

### Logistic regression for win probability: parameterization

I am modeling the outcome of baseball games using a logistic regression. I am struggling to understand the results of the analysis, and I believe this relates to the parameterization of the model. ...
• 709
23 views

### How to do prediction (evaluate marginal likelihood) in generative latent variable classifier?

The dataset is $\{\boldsymbol x_t,y_t\}$ for $t=1,\dots,T$, where $y_t \in \{0,1\}$. Define a generative latent variable classifier whose plate diagram is shown above. For each data point, a local ...
• 81
1 vote
My case is analyzing the association between the concentration of HIV DNA prior to therapy (time point 0, $t_0$) and the concentrations of biomarkers of HIV infection after therapy, measured in 6 time ...