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.

Filter by
Sorted by
Tagged with
0
votes
0answers
2 views

Comparison of forecasting models at scale

I am working on a project and 675 models time-series were built. Their variations are related with the intervals to be considered (1 year, 2 years or 9 years), the percentages of the training and test ...
0
votes
0answers
4 views

If you use a set of variables to calculate the target, can those variables be used as features?

For example, lets say I have some data on a set of products. This includes the cost of bringing the product to market and the total purchase amount of the product. Let's say I introduce a new variable ...
0
votes
0answers
7 views

Multi Target Techniques where Dependent Variables are Correlated

I have browser data that contains over 100 independent variables to predict customer spend. Instead of predicting total spend over a given time, let's say we want to predict the monthly spend for each ...
0
votes
1answer
39 views

Proving result about expectation under posterior predictive distribution

I am trying to figure out how to prove a result about the expectation of a random variable under the posterior predictive distribution, that may or may not be true. Let $X$ be a random variable, and $...
0
votes
0answers
9 views

Pearson's p value approach to test model's significance (pValueCompute.exe)

I have two questions regarding the calculation of p value and success rate, method stated in Pearson et al., 2007. Does this testing method work only when cross-validation is used in maxent setting? ...
0
votes
0answers
18 views

How do you write factor operations for rule-based conditional probability distributions?

Supposing I have rule-based conditional probability distributions (CPDs), $\{P(X|\text{Pa}_{X}), \cdots\}$, in a graphical model each represented as a set of rules $\mathscr{R}$ such that: For each ...
0
votes
0answers
30 views

Calculate forecast Arimax model step by step

I am making a model in R, Arima and I want to replicate the forecast in excel, but at the time of doing it I cannot continue due to the MA. this is just an example data to understand how the R ...
2
votes
1answer
11 views

How can I classify time-series given a predictor for each of them?

Say that I have two time-series and a predictor for each of them. I would like to build a classifer that given a window of future (and unseen) samples returns which series is more likely to have ...
1
vote
0answers
23 views

Forecasting monthly visitor count from daily values

I have a dataset of the daily visitor count of a website. Given this information, I want to forecast what the monthly visitor count will be. Depending on the visitor count on a day of the month, I ...
0
votes
0answers
18 views

How to normalize predicted values for an outcome event?

I work on predictive models for crime forecasting, meaning I try to model the risk for crimes. In the end of my modeling, I have the following values: number of predicted crimes for each state (...
2
votes
1answer
25 views

Using counterfactual modeling techniques to assess racial bias in predictive models

My team at a health insurance company is discussing how we might measure racial bias in the various predictive models our company uses to assess future health risk (such as annual medical cost or ...
0
votes
0answers
13 views

Correct loss function for this count statistics

I have time-series binned dataset, where each bin consists of count of particles. Since this is a physical phenomenon, I also have a model which predicts expected count of particles for each bin. ...
0
votes
0answers
8 views

Tag based recommendation system

Please bear with me cause I am not an expert in the subject. I want to build a recommendation system. It is not based on what others liked (collective based) or what I have liked (content based). The ...
-2
votes
0answers
11 views

Why is the predicted value rotated version of the actual value? [duplicate]

I am using random forest for regression analysis. Below is the plot of predicted value vs actual value. It seems that the predicted value is some rotated version of the actual value. Is there a reason ...
0
votes
0answers
12 views

Can I distiniguish impotant features in a dataset just by looking at the features and not at the target?

I have a matrix of examples X which contains only the features. I have a vector y which contains the target values (the things I'm trying to predict). In this case I'm trying to predict numerical ...
0
votes
0answers
27 views

Why is the predicted value rotated version of the actual value? [duplicate]

I am using random forest for regression analysis. Below is the plot of predicted value vs actual value. It seems that the predicted value is some rotated version of the actual value. Is there a reason ...
0
votes
0answers
7 views

Number of levels of categorical variable in Multilevel-Model

There is a rule of thumb that each prediction parameter in a regression must be supported by 10-15 observations. If I use dummy coding to represent the categorical variable in a multilevel-model, do ...
1
vote
0answers
12 views

Assessing results from package:IML FeatureImportance function output in R

Hello fellow CrossValidated friends! This question was migrated from StackOverflow, they said StatsCrossValidated was a good place to get an answer. I am trying to understand what it means in terms my ...
-1
votes
2answers
17 views

Another term for “real time decisioning”

There is another term for "real time decisioning," (RTD) and I need it to locate a paper (by searching). I've been beating my head over this all day. I'm thinking it's "[something] ...
1
vote
0answers
23 views

Predicting next destination (for a car)

I am trying to figure out how to predict the next destination of a an object (e.g. car) given its present location (and potentially last location), the type of object and potentially other factors as ...
1
vote
0answers
13 views

How to decide which pre-trained model to use for transfer learning?

For Deep Learning problems that deal with image data, how do I decide which pre-trained model architecture to use, like VGG or Resnet or Xception instead of trying them all(which will take days to ...
0
votes
1answer
92 views

chi square GLM inference

Suppose at $m$ different positions on a line $a_1,....,a_m$, we sample from a i.i.d normal distribution $N(\mu_i,\sigma_i^2)$, $n_i$ times for each of the $1\le i\le m$ different points. Here of ...
0
votes
0answers
24 views

Is there a method for evaluating predictive models that accounts for difference between predicted/actual distributions?

I am using machine learning to predict the monthly productivity (in dollars) of various groups of people. My question relates to ways of measuring the performance of my model. The distribution of ...
0
votes
0answers
19 views

Machine Learning for choosing the right cover box to contain smaller item boxes

I am studying on choosing the right box to contain item boxes. When customers order items, items have their own cases and ordered items are packed with the right cover box to contain them to deliver. ...
0
votes
0answers
9 views

How do I weight many-to-many relationships in a regression model?

In this context, I have 3 types of entities used to build a predictive model: subject - An individual. event - An activity performed by an individual at a particular time. target - An outcome that ...
0
votes
0answers
14 views

Interpreting GLM coefficients as multiplicative adjustments

I'm currently reviewing the provided solution for a GLM problem and I'm completely confused by the answers. The training data is staged as such: ...
0
votes
1answer
17 views

Advice on what type of model should be considered for treatment data

I have a dataset with four variables: treatment, subjectID, tumor size, and time. There are four different treatments with 10 subjects assigned to each treatment. The subjectID denotes which subject ...
1
vote
0answers
18 views

Need Concrete Example of Where a Frequentist Clustering Technique Outperforms the Existing Bayesian Clustering Techniques

I'm looking for a concrete example of frequentist clustering outperforming Bayesian clustering (using the best Bayesian algorithm for the problem, using the testing criteria below). There are many ...
0
votes
1answer
49 views

predictive distribution of beta [closed]

Is there a way of obtaning some kind of predictive posterior distribution for the beta model?
0
votes
0answers
10 views

Does ordinary kriging assumes a random field with zero drift?

The question is actually the above one. I'm a beginner in spatial stats and just trying to fit many puzzle pieces together. So I just wondered if ordinary kriging, in contrast to universal kriging, ...
0
votes
0answers
17 views

Survival analysis using glm and predict.glm on multiple input genes

I wonder if the current model is correct. What I am doing is combining two genes as input in a glm model that further is used to predict survival using survivalROC. ...
0
votes
0answers
21 views

VAR model is predicting better one period ahead

I am using a VAR(1) with two variables but it seems that the data forecast better one period ahead(blue line is the forecast, black line is the true value). Do you know why this happens? the data is ...
0
votes
0answers
12 views

Compute Tennis Match Winning Probabilities with Limited Predictors

I'm trying to predict tennis match outcomes in R using only match scores from previous opponents to predict winners/losers for future matches. The match scores, however, are only as granular as the ...
0
votes
1answer
47 views

Time Series Analysis vs Linear Regression for Stock Price Prediction

I see many online blogs that use Linear Regression to predict future stock price. I too have done this and my x variable is time elapsed. But I have been advised this problem is better suited for Time ...
0
votes
0answers
11 views

How to predict values for each country in a particular year?

I have a dataset that includes an outcome score for each country for over 4 years. The dataset also contains various predictors that could be used to predict the outcome scores. I want to build a ...
0
votes
1answer
66 views

Forecasting Sales LSTM - cannot capture peak values

I am trying to forecast retail sales for a company that have different stores. I currently use LSTM model which is designed as follow : data includes the info about sales between 2014-2020. After ...
1
vote
0answers
19 views

Forecast horizon and small sample problem

I have a stationary time series with slight persistence, if any; like stock returns. They are collected in weekly frequency, $N$ obs. I try to predict them with an $AR(1)$ model. The well known ...
1
vote
0answers
24 views

Addressing non-linearity in a fitted vs residuals plot

I am trying to conduct a data analysis project, which involves a multivariable regression model with 13 predictor variables. Before having transformed/ altered the data at all, I fitted a rough model ...
1
vote
1answer
28 views

How to Increase the weight for one predictor variable?

Just read this news: "Zuckerberg agreed to increase the weight that Facebook’s algorithm gave to NEQ scores to make sure authoritative news appeared more prominently." My question is: how ...
0
votes
0answers
7 views

How can I regularize data pre-processing parameter estimates (center, scale, rotation matrix)?

Suppose I have an $N×1$ vector $Y_{in}$ of response values an $N×P$ matrix of predictors $X_{in}$ whose individual columns exhibit significant correlation. Let's further suppose that this matrix $X$ ...
0
votes
0answers
13 views

Modeling additional structure within a distribution

Let's say the distribution of anthill locations in an area follows $\mu_1 \cdots \mu_k \sim \mathcal N (\mu_0, \Sigma_0)$. Additionally assume the distribution of ant locations in the anthill follows $...
1
vote
0answers
50 views

How to use causality knowledge to improve linear regression? [closed]

As Peter mentioned in this reply https://stats.stackexchange.com/a/26412/152503 Sometimes we do have a priori information about the causality relationship between features/predictors Using the fire ...
0
votes
1answer
32 views

Prediction of Variance vs Variance of Prediction

Is the prediction of the variance the same as the variance of the prediction? I know the concept of prediction intervals, used to specificy the variance of the prediction. I also know (G)ARCH models, ...
1
vote
0answers
68 views

ARMA models and predictive learning

It is well known that, in Time Series, ARMA models was proposed for make forecast about some $y_t$ based on past values of the process itself. However this is not the only possible way for make ...
2
votes
0answers
19 views

Continuous-time, non-recursive ARIMA Equation

In this question, I asked about validating the assumption of geometric Brownian motion in a analytic model using ARIMA. Here, I want to generalise this idea. If I'm building a decision model that ...
0
votes
0answers
20 views

What is the best approach to predict/ forecast the sales of 100s of parts?

I have been working on a sales database, aggregated and organized month wise, which has the sales trend of 600+ parts, of which 150 are major contributors. The parts can be aggregated into a part ...
0
votes
0answers
13 views

Explore hyperparameters for one model or try different models first?

When first approaching a regression problem, would you rather try a model you know well and explore its hyper parameters via grid-search + cross-validation, or would you instead explore different ...
0
votes
0answers
30 views

LDA vs multinomial logistic regression?

I'm trying to determine whether it is better to use multinomial logistic regression or a linear discriminant analysis to answer the question of which predictor variable more accurately predicts a ...
0
votes
0answers
10 views

Determining predictive power of different data types

I want to find out what the best way of predicting a categorical variable is. I have multivariate continuous data (positive and negative values) and another single categorical variable with several ...
1
vote
0answers
19 views

What metrics should I use when developing the first model?

When I build the first regression model for a new project, standard metrics such as MSE, MAE, R2 score can't tell if the first model will work well in production because there are no existing models ...

1
2 3 4 5
46