### Archivo

Archivo para la categoría ‘Statistical Modelling’

## Hierarchical forecasting of hospital admissions- ML approach (modeltime package)

The aim of this series of blog is to predict monthly admissions to Singapore public acute adult hospitals. EDA for the dataset was explored in past posts ( part 1 ; part 2 ).

library(tidyverse)
library(tidymodels)
library(timetk)
library(modeltime)
library(modeltime.ensemble)

# dataset and dataset for future forecast dats and pre-processing recipes were done in the past post. The output was uploaded onto my github.
url_datasets<-url("https://github.com/notast/hierarchical-forecasting/blob/main/4Dataset_ML.rds?raw=true")
close(url_datasets)

head(to_train,10)
....

link: https://notast.netlify.app/post/2021-06-14-hierarchical-forecasting-of-hospital-admissions-ml-approach-modeltime-package/

Categories: Statistical Modelling Tags:

Miércoles, 16 de junio de 2021 Sin comentarios

## An Alternative to the Correlation Coefficient That Works For Numeric and Categorical Variables

When starting to work with a new dataset, it is useful to quickly pinpoint which pairs of variables appear to be strongly related. It helps you spot data issues, make better modeling decisions, and ultimately arrive at better answers.

The correlation coefficient is used widely for this purpose, but it is well-known that it can’t detect non-linear relationships. Take a look at this scatterplot of two variables xx and yy.

Categories: Statistical Modelling Tags:

Sábado, 17 de abril de 2021 Sin comentarios

## Time Series Forecasting with XGBoost and Feature Importance

From https://datageeek.com/

Those who follow my articles know that trying to predict gold prices has become an obsession for me these days. And I am also wondering which factors affect the prices. For the gold prices per gram in Turkey, are told that two factors determine the results: USA prices per ounce and exchange rate for the dollar and the Turkish lira. Let’s check this perception, but first, we need an algorithm for this.

Categories: Statistical Modelling Tags:

Miércoles, 3 de marzo de 2021 Sin comentarios

## SEM Time Series Modeling

by Andrea Luciani

Andrea Luciani is a Technical Advisor for the Directorate General for Economics, Statistics and Research at the Bank of Italy, and co-author of the bimets package.

Structural Equation Models (SEM), which are common in many economic modeling efforts, require fitting and simulating whole system of equations where each equation may depend on the results of other equations. Moreover, they often require combining time series and regression equations in ways that are well beyond what the ts() and lm() functions were designed to do. For example, one might want to account for an error auto-correlation of some degree in the regression, or force linear restrictions modeling coefficients.

https://rviews.rstudio.com/2021/01/22/sem-time-series-modeling/

Categories: Statistical Modelling Tags:

Domingo, 24 de enero de 2021 Sin comentarios

## R-INLA review (part II)

The first review was about the main ideas and the approximations itself. This time its a review about the spatial models, see the new  arxiv’ed report.

Categories: Statistical Modelling Tags:

Viernes, 23 de febrero de 2018 Sin comentarios

## Probable Points and Credible Intervals

A continuación os enlazo dos «entradas de blog» muy muy interesantes:

Se trata de una introducción al la Teoría de Decisión desde la perspectiva  Bayesiana:

http://www.sumsar.net/blog/2014/10/probable-points-and-credible-intervals-part-one/

http://www.sumsar.net/blog/2015/01/probable-points-and-credible-intervals-part-two/

Categories: Statistical Modelling Tags:

Jueves, 8 de enero de 2015 Sin comentarios

## future of computational statistics

(This article was first published on Xi’an’s Og » R, and kindly contributed to R-bloggers)

I am currently preparing a survey paper on the present state of computational statistics, reflecting on the massive evolution of the field since my early Monte Carlo simulations on an Apple //e, which would take a few days to return a curve of approximate expected squared error losses… It seems to me that MCMC is attracting more attention nowadays than in the past decade, both because of methodological advances linked with better theoretical tools, as for instance in the handling of stochastic processes, and because of new forays in accelerated computing via parallel and cloud computing, The breadth and quality of talks at MCMski IV is testimony to this. A second trend that is not unrelated to the first one is the development of new and the rehabilitation of older techniques to handle complex models by approximations, witness ABCExpectation-Propagation, variational Bayes, &tc. With a corollary being an healthy questioning of the models themselves. As illustrated for instance in Chris Holmes’ talk last week. While those simplifications are inevitable when faced with hardly imaginable levels of complexity, I still remain confident about the “inevitability” of turning statistics into an “optimize+penalize” tunnel vision…  A third characteristic is the emergence of new languages and meta-languages intended to handle complexity both of problems and of solutions towards a wider audience of users. STAN obviously comes to mind. And JAGS. But it may be that another scale of language is now required…

If you have any suggestion of novel directions in computational statistics or instead of dead ends, I would be most interested in hearing them! So please do comment or send emails to my gmail address bayesianstatistics…

Categories: Statistical Modelling Tags:

Lunes, 29 de septiembre de 2014 Sin comentarios

Interesante post sobre la interacción R y  twitter.

# Celebrity twitter followers by gender

May 25, 2014
_______________

Categories: Statistical Modelling Tags:

Lunes, 26 de mayo de 2014 Sin comentarios

## Los mejores trabajos del 2014

ALICIA NIETO | 16 abril de 2014 |

Vamos a ayudarte a contestar a una de las preguntas más difíciles de hacer frente a lo largo de nuestra vida: «¿en qué quiero trabajar?». La web especializada en búsqueda de empleoCareerCast se ha encargado de hacernos más llevadero el trabajo y ha seleccionado los 10 mejores trabajos del 2014. Para llevar a cabo el ranking ha evaluado diferentes características relacionadas con 200 trabajos diferentes, como los ingresos, el esfuerzo, el ambiente laboral o las opciones de futuro. La lista definitiva es la siguiente:

1. Matemático: salario medio anual de $101.360 y crecimiento de empleo previsto para el 2022 del 23%. 2. Profesor de Universidad: salario medio anual de$68.970 y crecimiento de empleo previsto para el 2022 del 19%.

3. Estadístico: salario medio anual de $75.560 y crecimiento de empleo previsto para el 2022 del 27%. 4. Experto en seguros: salario medio anual de$93.680 y crecimiento de empleo previsto para el 2022 del 26%.

5. Auditólogo: salario medio anual de $69.720 y crecimiento de empleo previsto para el 2022 del 34%. 6. Higienista dental: salario medio anual de$70.210 y crecimiento de empleo previsto para el 2022 del 33%.

7. Ingeniero de software: salario medio anual de $93.350 y crecimiento de empleo previsto para el 2022 del 22%. 8. Analista de sistemas informáticos: salario medio anual de$79.680 y crecimiento de empleo previsto para el 2022 del 25%.

9. Terapeuta ocupacional: salario medio anual de $75.400 y crecimiento de empleo previsto para el 2022 del 29%. 10. Patólogo del habla: salario medio anual de$69.870 y crecimiento de empleo previsto para el 2022 del 19%.

Categories: Statistical Modelling Tags:

Martes, 29 de abril de 2014 Sin comentarios