Category: R

Feed: R-bloggers. Author: sang-heon lee. #========================================================## Quantitative ALM, Financial Econometrics & Derivatives # ML/DL using R, Python, Tensorflow by Sang-Heon Lee ## https://kiandlee.blogspot.com#——————————————————–## Estimation of Nelson-Siegel-Svensson model#========================================================#graphics.off(); rm(list = ls())library(alabama)#———————————————–# Nelson-Siegel-Svensson function#———————————————–# NSS factor loadingnss_loading – function(la,m) {    la1 – la[1]; la2 – la[2]    C  – cbind(        rep(1,length(m)),            (1–exp(–la1*m))/(la1*m),                     (1–exp(–la1*m))/(la1*m)–exp(–la1*m),        (1–exp(–la2*m))/(la2*m)–exp(–la2*m))    return(C)}# fitting functionnss_fit – function(para, m) {    beta – para[1:4]    return(nss_loading(para[5:6],m)%*%beta)}# objective functionnss_objf – function(para, y, m) {    return(sqrt(mean((y – nss_fit(para, m))^2)))}# constrOptim.nl constraint function ( > )nss_constf – function(para, y, m) {    beta – para[1:4]     la1  – para[5]; la2 – para[6]    h    – rep(NA, 6)    # b1 > 0, b1+b2 > 0    h[1] – beta[1]             h[2] – beta[1]+beta[2]    # 0.15 > la1 > 0.03    h[3] – 0.15–la1    h[4] – la1–0.03    # 0.025 > la2 > 0.0075    h[5] – 0.025–la2    h[6] – la2–0.0075    return(h)}#===========================================================# 1. Read data#===========================================================# Estimated parameters of Nelson-Siegel (Benchmark)str.ns_est– “    beta1      beta2       beta3     lambda       rmse    7.949446  0.2933681 -0.08749462 0.11555381 0.07665367    7.052864 -1.5778886 -0.42253575 0.01823114 0.06432032    5.586903 -0.5600670 -1.43976821 0.03357762 0.06423196    6.011739  0.3470154 -1.00018107 0.04565435 0.09314726″m.ns_est – read.table(text = str.ns_est, header=TRUE)str.zero – “    mat 19890630 19950929 19980831 20000929    3   8.171964 5.413123 4.931036 6.180960    6   8.143305 5.509881 4.949141 6.207158    9   8.123782 5.592753 4.962873 6.204367    12  8.111922 5.654558 4.962513 6.176521    24  7.974091 5.782513 4.824109 5.932045    36  7.943757 5.841313 4.868401 5.862433    48  7.959505 5.920732 4.888494 5.825789    60  7.942511 5.967207 4.900205 5.799035    72  8.000704 6.013260 4.980854 5.826354    84  8.010956 6.055806 4.977891 5.847383    96  8.020059 6.151399 5.007556 5.859937    108 8.027901 6.224049 5.054961 5.875451    120 8.037343 6.265079 5.100250 5.912506    144 8.063573 6.364383 5.200412 6.002328    180 8.052892 6.512078 5.347487 6.093959    240 7.971860 6.718515 5.482335 6.091155    300 7.861711 6.789321 5.422557 5.972525    360 7.721083 6.599990 5.246325 5.697709″df – read.table(text = str.zero, header=TRUE)m  – df$mat; nmat – length(m)#===========================================================# 2. Parameter estimation#===========================================================ctrl.optim – list(maxit=50000, trace=0,                    reltol = 1e–10, abstol = 1e–10)ctrl.outer – list(eps = 1e–10, itmax = 50000,                    method = “Nelder-Mead”, NMinit = TRUE)# output containerm.nss_est – matrix(NA, 4, 7)m.nss_fit – matrix(NA, length(m), 4)# Estimation of Nelson-Siegel-Svenssonfor(i in 1:4) {    y – df[,1+i]    x_init – c(y[nmat], y[1]–y[nmat],                 2*y[6] –y[1]–y[nmat],                 2*y[15]–y[1]–y[nmat], 0.0609, 0.01)    opt – constrOptim.nl(y = y, m = m,        par=x_init, fn=nss_objf, hin=nss_constf,        control.optim =ctrl.optim,        control.outer=ctrl.outer)     m.nss_est[i,] – c(opt$par, opt$value)    m.nss_fit[,i] – nss_fit(opt$par, m)    colnames(m.nss_est) – c(“beta1”, “beta2”,         “beta3”, “beta4”, “lambda1”, “lambda2”,“rmse”)}colnames(m.nss_fit) – paste0(“fitting-“,1:4)#=======================================================# 3. Estimation results#=======================================================# NSS factor loadingmi – 1:360x11(width = 16/2.5, height = 5); matplot(mi, nss_loading(c(0.0609, 0.01),mi),         type = “l”, lwd=6, xlab = “Maturity (month)”,        ylab = “Factor loading”)legend(“right”,        legend=c(“Level”, “Slope”, “Curvature1”, “Curvature2”),       pch = c(15,16), border=“white”, box.lty=0, cex=1,       col = c(“black”,“#DF536B”, “#61D04F”, “#2297E6”))# Parameter estimatesm.ns_estm.nss_est# data and fitted yieldsround(cbind(df[,–1], m.nss_fit),3) ... Read More

Feed: R-bloggers. Author: Achim Zeileis. [This article was first published on Achim Zeileis, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here) Want to share your content on R-bloggers? click here if you have a blog, or here if you don't. A new arXiv paper investigates which building blocks of random forests, especially causal forests and model-based forests, make them work for heterogeneous treatment effect estimation, both in randomized trials and observational studies. Citation Susanne Dandl, Torsten Hothorn, Heidi Seibold, Erik Sverdrup, Stefan Wager, Achim Zeileis (2022). “What Makes Forest-Based Heterogeneous Treatment Effect ... Read More

Feed: R-bloggers. Author: kjytay. I recently noticed a slight inconsistency between the forcats package’s fct_lump_min and fct_lump_prop functions. (I’m working with v0.5.1, which is the latest version at the time of writing.) These functions lump levels that meet a certain criteria into an “other” level. According to the documentation, fct_lump_min “lumps levels that appear fewer than min times”, and fct_lump_prop “lumps levels that appear fewer than prop * n times”, where n is the length of the factor variable. Let’s try this out in an example: x <- factor(c(rep(1, 6), rep(2, 3), rep(3, 1))) x # [1] 1 1 1 1 ... Read More

Feed: R-bloggers. Author: matloff. [This article was first published on Mad (Data) Scientist, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here) Want to share your content on R-bloggers? click here if you have a blog, or here if you don't. Prominent statistician Frank Harrell has come out with a radically new R tutorial, rflow. The name is short for “R workflow,” but I call it “R in a box” –everything one needs for beginning serious usage of R, starting from little or no background. By serious usage I mean real applications in ... Read More

Feed: R-bloggers. Author: Stéphane Laurent. Given a number (s geqslant 1), the following map: [ (u, v) mapsto (x, y, z) = frac{Bigl(scosfrac{u}{s}, ssinfrac{u}{s}, sin vBigr)}{sqrt{s^2+1}-cos v} ] is a conformal parameterization of the torus (the donut), where (-spi leqslant u and (pi leqslant v . I found it in this paper by J.M. Sullivan. The number (s) is the ratio of the major radius over the minor radius. The conformality of the map has the following consequence: you can easily map a doubly periodic image on the torus in such a way that it will perfectly fit on the ... Read More

Feed: R-bloggers. Author: R on head spin - the Heads or Tails blog. Building a Kaggle Leaderboard with gtExtras | R-bloggers % select(rank, avatar, author, total, perc) %>% gt() %>% gt_img_rows(avatar, img_source = "local") %>% gt_plt_bar_pct(column = perc, scaled = TRUE, fill = kaggle_blue) %>% gt_highlight_rows(rows = 1, font_weight = "normal", fill = kaggle_gold) %>% gt_highlight_rows(rows = 2, font_weight = "normal", fill = kaggle_silver) %>% gt_highlight_rows(rows = 3, font_weight = "normal", fill = kaggle_bronze) %>% tab_header( title = md("**Kaggle Hidden Gems Competition - Final Results**"), subtitle = html(str_c("A 2022 Notebooks Competition scored by Community Judges ")) ) %>% tab_style( style ... Read More

Feed: R-bloggers. Author: The Jumping Rivers Blog. Introduction RStudio Workbench provides a development environment for R, Python, and many other languages. When developing a performant web application you may progress from Shiny towards tools like Plumber. This allows you to continue development of the true application code, modelling, data processing in the language you already know, R, while providing an interface to those processes suitable for embedding in larger web-based applications. As a Shiny developer, one popular front-end library you might already be familiar with is React. React is used in many popular {htmlwidgets} such as {reactable} and anything built ... Read More

Feed: R-bloggers. Author: Selcuk Disci. [This article was first published on DataGeeek, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here) Want to share your content on R-bloggers? click here if you have a blog, or here if you don't. The war between Russia and Ukraine has affected the global food supply other than many vital things. Primarily cereal crop products have been affected the most because the imports have been provided to the world mainly through Ukraine and Russia. Let’s check the situation we’ve mentioned for G20 countries. library(tidyverse) library(fpp3) library(bbplot) library(plotly) ... Read More

Feed: R-bloggers. Author: Benjamin Smith. [This article was first published on r – bensstats, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here) Want to share your content on R-bloggers? click here if you have a blog, or here if you don't. Natural Language Processing (NLP) is a powerful tool in the Machine Learning landscape that can (among other things) allow users to classify sentiment and predict text. Many of recent my blogs have been about data manipulation and data engineering, so I decided change things up to look into showing some applications ... Read More

Feed: R-bloggers. Author: Jim. [This article was first published on Data Science Tutorials, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here) Want to share your content on R-bloggers? click here if you have a blog, or here if you don't. The post How to Join Multiple Data Frames in R appeared first on Data Science Tutorials How to Join Multiple Data Frames in R?, you can find it useful to connect many data frames in R. Fortunately, the left join() function from the dplyr package makes this simple to accomplish. Crosstab calculation ... Read More