Publications by H. K. Tseng
Statistical Learning: Week-13
Leave-One-Out cross validation (LOOCV) Model validation using LOOCV # load required packages and data (we will be using AmesHousing data) # install.packages(c("boot", "rsample", "AmesHousing")) library(ISLR) library(caret) ## Loading required package: ggplot2 ## Loading required package: lattice library(rsample) library(boot) ## ## Attaching ...
379 sym R (17713 sym/59 pcs) 1 img
Statistical Learning: Week-12
\(K\)-means and hierarchical clustering for wine types Apply \(K\)-means and hierarchical clustering to wine data # load required packages and data (wine data) install.packages(c("HDclassif", "useful", "factoextra")) ## Installing packages into 'C:/Users/hktse/Documents/R/win-library/3.6' ## (as 'lib' is unspecified) ## package 'HDclassif' succ...
326 sym R (15142 sym/117 pcs) 7 img
Statistical Learning: Week-11
Naïve Bayes ## load required packages and data library(dplyr) ## ## Attaching package: 'dplyr' ## The following objects are masked from 'package:stats': ## ## filter, lag ## The following objects are masked from 'package:base': ## ## intersect, setdiff, setequal, union library(ggplot2) library(tidyr) library(caret) ## Loading re...
695 sym R (59626 sym/249 pcs) 18 img
Statistical Learning: Week-10
Classifying Legendary Pokemons using SVM! Using SVM techniques to classify if a Pokemon is a legendary one. # install required packages in one-swoop install.packages(c("dplyr", "ggplot2", "tidyr", "reshape2", "caret", "skimr", "psych", "e1071", "data.table", "Matrix", "keras")) ## Installing packages into 'C:/Users/hktse/Documents/R/win-library/3....
487 sym R (36942 sym/145 pcs) 3 img
Understanding Support Hyperplane
Understanding separating hyperplane (or support hyperplane) # generate random data # random random seed for reproducibility set.seed(123) #set n = 500 data points. n <- 500 #Generate data frame with two uniformly distributed predictors lying between 0 and 1. df <- data.frame(x1 = runif(n), x2 = runif(n)) # create a va...
106 sym R (4806 sym/16 pcs) 4 img
Statistical Learning: Week-9
Ridge and LASSO regressions Fit ridge and LASSO regressions, interpret coefficients and visualize their variation across the range of \(\lambda\). # load the required packages and data library(glmnet) library(caret) library(plotmo) data(mtcars) names(mtcars) # as usual, check out what's inside the loaded dataframe ## [1] "mpg" "cyl" "disp"...
418 sym R (24994 sym/63 pcs) 4 img
Note: LASSO, Ridge, and Penalty explained
LASSO vs. Ridge LASSO and Ridge regression are regression methods that perform \(variable\) \(selection\) and \(regularization\) to enhance the prediction accuracy and interpretability of the statistical model. In short, they do two things Variable selection: identify important variables in the data that explain major variation in the outcome ...
4688 sym Python (10112 sym/41 pcs) 3 img
Statistical Learning: Week-8
set.seed(123) n = 10 xr = seq(0, n, by=.1) # generate a random data from a sin function plus some random errors yr = sin(xr/2) + rnorm(length(xr))/2 # combine x and y into a df for easy manipulation df = data.frame(x = xr, y = yr) # plot the data plot(df) lm.fit = lm(y ~ x, data = df) abline(lm.fit, col = "red") # If the degree of the p...
646 sym R (25161 sym/92 pcs) 20 img
Statistical Learning: Week-6
Binary outcome as a generalization of linear regression model with limited dependent variables Estimating a model of binary outcome using lm() function library(ISLR) data(Default) names(Default) ## [1] "default" "student" "balance" "income" ## convert the outcome variable "default" to numeric: Yes = 1, No = 0 # the ifelse logic: ifelse(if a var...
857 sym R (38844 sym/161 pcs) 7 img
Statistical Learning: Week-5
Non-normality test ## load Prestige data from the "car" package library(car) ## Loading required package: carData lm.fit <- lm(prestige ~ education, data = Prestige) plot(lm.fit) ## Shapiro test of normality shapiro.test(Prestige$prestige) ## ## Shapiro-Wilk normality test ## ## data: Prestige$prestige ## W = 0.97198, p-value = 0.02875 ...
1097 sym R (32547 sym/104 pcs) 15 img