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 ...

learn more… | top users | synonyms

1
vote
1answer
24 views

Confusion about Cross-Validation for Hierarchical Bayesian Regression Models

I had two questions regarding model selection for a Hierarchical Bayesian (HB) Regression Model and the purpose of Cross-Validation. 1). I understand cross-validation as one way to perform model ...
4
votes
2answers
62 views
+100

Statistical model to predict the next move on network only using movement history

Is it possible to build a statistical model that predicts the next move in a graph solely based on past movements and the structure of the graph? I have made an example to illustrate the problem: ...
5
votes
1answer
99 views

Goodness-of-fit test in Logistic regression; which 'fit' do we want to test?

I am referring to this question and its answers: How to compare (probability) predictive ability of models developed from logistic regression? by @Clark Chong and answers/comments by @Frank Harrell. ...
0
votes
0answers
18 views
+50

Can it be as accurate to model child-variables to estimate a parent-variable instead of modeling the parent-variable directly?

With time series data, let's say you want to model the return of the S&P 500. Could you get as good or better results by modeling each stock, and aggregating them to estimate the return of the ...
2
votes
1answer
58 views

Difference Between Discrete Time Proportional Hazards and Logistic Regression

My data consists of one row per person, per month that person was "exposed" to an event. So the month is the discrete time and the row corresponds to one "person-month". There are a few independent ...
0
votes
0answers
14 views

Calculating optimal alpha, beta, and gamma for Holt-Winters in R [closed]

Is there a way to calculate the alpha, beta, and gamma parameters in R for Holt-Winters? I tried using the one provided from the HoltWinters formula but it didn't give an accurate answer for ...
0
votes
1answer
14 views

Making a mob-like decision tree with pre-specified splits and models for leaves

I would like to make a special kind of hybrid tree model in R, similar to the mob models in the party and ...
0
votes
0answers
15 views

Predicting lat/long from binary features

I have a number of observations that occur around my city (a small area), and several of them have latitude and longitude. I have been looking into predicting the latitude/longitude of the ...
2
votes
0answers
35 views

Boosting: why is the learning rate considered a regularization parameter?

I understand that the learning rate parameter ($\nu \in [0,1]$) in Gradient Boosting shrinks the contribution of each new base model -typically small trees- that is added in the series. It was shown ...
0
votes
0answers
10 views

Predicting from individual and group data

Suppose that I have a classroom of students who each have taken two tests (A and B). I have the scores for test A and only know the average score for test B. Since students tend to perform similarly ...
1
vote
1answer
14 views

predict upvotes a review would get in a certain time period after publishing--yelp review

I've encountered this question online. How to design an algorithm to predict how many upvotes a review would get in a certain time period after publishing. Let's say the review comes from Yelp. So the ...
0
votes
0answers
11 views

Using one model to validate another? (Validation with no ground truth?)

My question is on model validation when there is no ground truth. Suppose I have a data set of attributes: $x = \{x_1,\ldots, x_n\}$ These attributes are different attributes that describe animals ...
0
votes
0answers
12 views

Relationship between predictive capability of a model and interpretability

I don't understand if there is a relationship between the predictive capability of a model and the possibility to use it for interpretation. Suppose I have fitted a linear model using only the ...
0
votes
0answers
27 views

How can I predict one time series using another time series?

Necessary Information: I have time series $X_t$ and $Y_t$ and $Z_t$, $t=0,...,N$. I want to develop a model to use $X_t$ to predict $Y_t$ where I know there exists a relationship $Y_t = Y_{t-1} + Z_t ...
0
votes
1answer
29 views

How good is a model if it can't predict a single positive class?

I have a training set of over a 100,000 points that is used to train a Logistic Regression Classifier (logit, since response is binary). The model is testing/fitted on a test set of 20,000 items. The ...
0
votes
1answer
31 views

Meaning of Dummy Variable

I am working on a case study from Kaggle: https://www.kaggle.com/c/liberty-mutual-fire-peril Here, in the variables, there is a variable: "Dummy" about which the description says: (dummy: Nuisance ...
0
votes
0answers
10 views

Using Model on less variables

Something I've been confused about for a while, If I built a predictive model on say, 10 variables and predicted a test set with that model, could I then use those predictions as a training set and ...
1
vote
1answer
48 views

prediction on short time series with seasonality and data correlations

I have, say, 5 weeks of data standing for daily income of a company and I want to predict the next income. Obviously, there is a seasonality in data - every day is "seasonal" with the same day of the ...
0
votes
0answers
18 views

Statistical/ML models when observations have different amounts of input

Let's say we're predicting an employee's performance review score for the following year based on his/her performance review scores from each previous year of their employment. We might have these ...
1
vote
0answers
23 views

How to account for high-score outliers in predictive sport metrics?

I'm currently trying to rate different teams in a league based on their underlying skill and eventually make predictions on future games using these simple ratings. I have so far found that a team ...
0
votes
0answers
19 views

Lasso Regression in SAS

I tried running the LASSO regression in SAS using the glmselect procedure. I started off with a set of 2000 variables and ended up with 37 after running the code. However, I feel that the final 37 ...
1
vote
1answer
55 views

Multiple (not independent) response variables in machine learning

Question: How to predict the percentage of people with age < 18, 18-65 and > 65 who visit a webpage using machine learning in R? Since these percentages sum to 100 for all observations, they are ...
4
votes
1answer
86 views

What's the reasoning behind presenting unvalidated AUC as a measure of model fit or performance?

Often one sees, particularly in the biomedical literature, papers that analyze the performance of a risk prediction model in terms of the AUC or the area under the ROC curve. If the AUC is suitably ...
2
votes
1answer
22 views

“Knot range too wide” problem when fitting a GAM model

This question may be very basic. I'm fitting a GAM model (for two-class classification) with a few numeric variables. Some of them are like this: a large portion of values are set at -999, meaning no ...
0
votes
1answer
18 views

Data analysis using Gross vs. Variation Data

I am trying to find a macroeconomic model that fits my data mainly using simple multiple regression. However, I am a little confused with the mistakes I may be doing. The question below might sound ...
0
votes
0answers
18 views

How to decide who gets an intervention?

Suppose we want to see whether a particular marketing strategy is effective at getting people to buy a product. We have historical data of people and how the marketing strategy affected their ...
3
votes
1answer
55 views

GARCH vs SV for Forecasting

I believe I am aware of how GARCH family and stochastic volatility models differ in their construction and assumptions on the volatility states, (i.e. GARCH family assumes deterministic volatility ...
0
votes
1answer
53 views

R train random forest for positive or negative predicitve value, not accuracy

I am working with random forests on financial data (predicting if stock rises versus falls). I figured out that I get better performance, if I build one model for "rising" and one for "falling". ...
1
vote
0answers
12 views

AIC and categorical predictors

As we know, the AIC is defined as $\mathrm{AIC} = 2k - 2\ln(L)$ with $k$ being the amount of estimated parameters. My question is: in case I have a categorical predictor, lets say: educational ...
4
votes
2answers
111 views

How to compare (probability) predictive ability of models developed from logistic regression?

I know some well-known measures are $c$ statistic, Kolmogorov-Smirnov $D$ statistic. However, as far as I know, those statistics take into account only of the rank order of the observations, and is ...
0
votes
1answer
21 views

Do models with multiple covariates and single covariate models differ from each other?

I have a binary response variable with 9 predictor variables. Lets denote the predictors $A, B, C, D, E...$ Suppose I run a model $y_i = \beta_0 + \beta_1 A + \beta_2 B + \beta_3 C + \cdots$. The ...
0
votes
0answers
11 views

Interpretation of standardized beta coefficient estimates and use within the exponential formula for prediction purposes

I'm working on a data set where I plan to use logistic regression to evaluate non-random habitat selection for a wildlife species. My dependent variable is 1 = used location by an animal and 0 = ...
1
vote
0answers
31 views

Root-Mean Squared Error for Bayesian Regression Models

I'm trying to get a sense of my prediction errors for a Bayesian regression model and I was using the Root-Mean-Squared Error. My question is, since are predictions are stochastic, would it make ...
0
votes
1answer
34 views

Differences between cross validation and bootstrapping to estimate the standard error of the AUC of a given ROC curve

I know there's been some discussion on differences between CV and bootstrapping for estimating out-of-sample prediction error of a classifier. For example, in here (Differences between cross ...
20
votes
6answers
3k views

Should parsimony really still be the gold standard?

Just a thought: Parsimonious models have always been the default go-to in model selection, but to what degree is this approach outdated? I'm curious about how much our tendency toward parsimony is a ...
0
votes
0answers
22 views

Use Logistic Regression Literature for Logit Discrete Choice Models

I'm currently developing a binary logit Discrete Choice Model (DCM) in the context of my thesis. Obviously, I want to develop the model following academic standards. A few questions have been arising: ...
23
votes
9answers
2k views

When can correlation be useful without causation?

A pet saying of many statisticians is "Correlation doesn't imply causation." This is certainly true, but one thing that DOES seem implied here is that correlation has little or no value. Is this ...
0
votes
1answer
29 views

Appropriate predictive model for two random time series with serial correlation

Say I have annual observations of the temperatures at the North Pole and South Pole for many years. I want to build a model that given the South Pole temperature for the current year and all prior ...
2
votes
2answers
99 views

Random Forest - Numeric and Dummy Variables together

I am trying to create a logistic regression model and a random forest model on the same data to predict probability of default. For the logistic regression model, I have created some dummy variables ...
0
votes
0answers
14 views

Predicting responses to test-data variables that lay far outside model training-data variable range

What can one do to improve model predicting accuracy when using test data that has a variable with magnitudes far outside the range of that variable in the training data? Example: This question ...
2
votes
1answer
17 views

Methodological test for choosing 'worse' models that make 'better' (more realistic) predictions?

I've run 4 models (simple LM, quadratic model, GLMM, and GLMM with quadratic) to predict tree age (age) from tree diameter (D) for each of 42 species (SPEC). The diameter data has all been log ...
0
votes
0answers
14 views

Interpolation of Data Value using Optimized Weighting of Its Features

I have a question regarding "Interpolation" / "Prediction" of a value. Assuming we have a data set $ { \left\{ \left( {x}_{i}, {y}_{i} \right) \right\}}_{i = 1}^{N} $ where $ {x}_{i} \in ...
0
votes
1answer
40 views

Regression model for edge-sensitive data set

I have data sets in which important information is allocated in the edges, which are also very sensitive to inaccuracies. I would like to find a regression model based on edge recognition that brings ...
4
votes
1answer
125 views

Relative variable importance for Boosting

I'm looking for an explanation of how relative variable importance is computed in Gradient Boosted Trees that is not overly general/simplistic like: The measures are based on the number of times a ...
1
vote
1answer
32 views

Beyond least squares: how to choose a predictive model or algorithm? (reference request)

There are dozens of algorithms one can use to build a predictive model. What books or studies exist that can help one determine which algorithm to use? Elements of Statistical Learning spends a lot ...
3
votes
1answer
41 views

Forecasting monthly time series with known periodicity and a known driver

For 2004-2014, I have monthly measurements of my outcome of interest - some kind of physical exposure - for a collective of many thousand persons. The main determinant for the average exposure level ...
0
votes
0answers
33 views

Building a Predictive Model

I'm inexperienced and confused in statistics, so I need help. I have a data table, values are temperature, particulate matter(PM), and vegetation indexes. And idea is that when PM increases, ...
1
vote
1answer
77 views

Bayesian approach for comparing the predictability of different datasets for another

Suppose I have three datasets A, B and C with not necessarily the same amount of data. Now, I want to know whether dataset A or dataset B is better in predicting C. I want to use a Bayesian approach ...
0
votes
1answer
9 views

Unable to predict using bart() {BayesTree}

I used bart function from BayesTree library to build a model on my training data. It fits my training data very well. However, I'm unable to predict for the test set and check its performance. ...
1
vote
0answers
35 views

Regression to Predict a 'Discrete' Ratio

I am to build a model $y_i \sim f(X_i, n_i) + error_i$. The regressor $y_i \in [0, n_i]$. $y_i$ and $n_i$ are positive integers. Each observation $i$ has a different known $n_i$, and $n_i$ varies ...