library(quantmod)
## Warning: package 'quantmod' was built under R version 4.3.3
## Loading required package: xts
## Warning: package 'xts' was built under R version 4.3.3
## Loading required package: zoo
## Warning: package 'zoo' was built under R version 4.3.3
##
## Attaching package: 'zoo'
## The following objects are masked from 'package:base':
##
## as.Date, as.Date.numeric
## Loading required package: TTR
## Warning: package 'TTR' was built under R version 4.3.3
## Registered S3 method overwritten by 'quantmod':
## method from
## as.zoo.data.frame zoo
library(PerformanceAnalytics)
## Warning: package 'PerformanceAnalytics' was built under R version 4.3.3
##
## Attaching package: 'PerformanceAnalytics'
## The following object is masked from 'package:graphics':
##
## legend
library(tidyverse)
## Warning: package 'tidyverse' was built under R version 4.3.3
## Warning: package 'readr' was built under R version 4.3.3
## Warning: package 'dplyr' was built under R version 4.3.3
## Warning: package 'lubridate' was built under R version 4.3.3
## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr 1.1.4 ✔ readr 2.1.5
## ✔ forcats 1.0.0 ✔ stringr 1.5.1
## ✔ ggplot2 3.5.2 ✔ tibble 3.2.1
## ✔ lubridate 1.9.3 ✔ tidyr 1.3.0
## ✔ purrr 1.0.2
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::first() masks xts::first()
## ✖ dplyr::lag() masks stats::lag()
## ✖ dplyr::last() masks xts::last()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors
library(PortfolioAnalytics)
## Warning: package 'PortfolioAnalytics' was built under R version 4.3.3
## Loading required package: foreach
## Warning: package 'foreach' was built under R version 4.3.3
##
## Attaching package: 'foreach'
##
## The following objects are masked from 'package:purrr':
##
## accumulate, when
##
## Registered S3 method overwritten by 'PortfolioAnalytics':
## method from
## print.constraint ROI
library(knitr)
library(readxl)
## Warning: package 'readxl' was built under R version 4.3.3
library(ROI)
## Warning: package 'ROI' was built under R version 4.3.3
## ROI: R Optimization Infrastructure
## Registered solver plugins: nlminb, symphony, quadprog.
## Default solver: auto.
##
## Attaching package: 'ROI'
##
## The following objects are masked from 'package:PortfolioAnalytics':
##
## is.constraint, objective
library(ROI.plugin.quadprog)
## Warning: package 'ROI.plugin.quadprog' was built under R version 4.3.3
library(tseries)
## Warning: package 'tseries' was built under R version 4.3.3
library(Matrix) # nearPD
##
## Attaching package: 'Matrix'
##
## The following objects are masked from 'package:tidyr':
##
## expand, pack, unpack
# Đường dẫn file của bạn (sửa lại nếu cần)
file_path <- "C:/Users/ASUS/Downloads/data 2.xlsx"
# Đọc dữ liệu
raw_df <- read_excel(file_path, sheet = 1)
# Đổi tên cột về định dạng nhất quán (phòng khi tên cột viết hoa/thường khác nhau)
names(raw_df) <- toupper(names(raw_df))
# Kiểm tra bắt buộc các cột cần có
required_cols <- c("DATE","HPG","MWG","FPT","VCB","VNM","VN30")
stopifnot(all(required_cols %in% names(raw_df)))
# Chuẩn hoá cột DATE
if (!inherits(raw_df$DATE, "Date")) {
# Nếu cần, bạn có thể chỉ định định dạng ngày cụ thể, ví dụ: "%Y-%m-%d" hay "%d/%m/%Y"
raw_df$DATE <- as.Date(raw_df$DATE)
}
# Hàm chuyển các cột ký tự có dấu phẩy trong số sang numeric
to_numeric_clean <- function(x) {
if (is.character(x)) {
as.numeric(gsub(",", "", x))
} else {
as.numeric(x)
}
}
# Chuẩn hoá toàn bộ các cột giá (trừ DATE)
price_df <- raw_df %>%
mutate(across(all_of(setdiff(required_cols, "DATE")), to_numeric_clean)) %>%
arrange(DATE)
# Chuyển sang xts
prices <- xts(price_df[, setdiff(required_cols, "DATE")], order.by = price_df$DATE)
# Xem nhanh
kable(head(prices), caption = "Dữ liệu giá sau khi chuẩn hoá")
Dữ liệu giá sau khi chuẩn hoá
| 3459.4 |
9.957 |
7437.9 |
16.822 |
66667 |
599.69 |
| 3492.0 |
9.957 |
7484.6 |
17.929 |
66667 |
606.37 |
| 3459.4 |
9.773 |
7531.3 |
17.718 |
68056 |
607.65 |
| 3459.4 |
9.865 |
7453.5 |
18.245 |
68403 |
606.85 |
| 3524.7 |
10.049 |
7515.7 |
19.458 |
71528 |
618.41 |
| 3524.7 |
10.142 |
7562.4 |
19.616 |
69444 |
615.73 |
Thiết lập tham số &
tính log-returns
# Tham số
market_index <- "VN30"
tickers <- setdiff(colnames(prices), market_index)
# Lãi suất phi rủi ro (có thể chỉnh)
rf_annual <- 0.045
rf_daily <- (1 + rf_annual)^(1/252) - 1
# Log-returns (bỏ hàng NA đầu)
library(PerformanceAnalytics)
ret_all <- Return.calculate(prices, method = "log")
ret_all <- ret_all[-1, ]
stock_returns <- ret_all[, tickers, drop = FALSE]
market_returns <- ret_all[, market_index, drop = FALSE]
# Kiểm tra nhanh
kable(head(stock_returns), caption = "Log-returns cổ phiếu (sample)")
Log-returns cổ phiếu (sample)
| 0.009379474 |
0.000000000 |
0.006259025 |
0.063731960 |
0.000000000 |
| -0.009379474 |
-0.018652340 |
0.006220093 |
-0.011838442 |
0.020620818 |
| 0.000000000 |
0.009369658 |
-0.010383949 |
0.029309998 |
0.005085788 |
| 0.018700163 |
0.018479988 |
0.008310445 |
0.064367227 |
0.044672298 |
| 0.000000000 |
0.009212090 |
0.006194434 |
0.008087263 |
-0.029568309 |
| 0.000000000 |
0.044451763 |
0.002060712 |
-0.005367151 |
0.019804321 |
kable(head(market_returns), caption = "Log-returns thị trường (VN30)")
Log-returns thị trường (VN30)
| 0.011077506 |
| 0.002108698 |
| -0.001317415 |
| 0.018870024 |
| -0.004343112 |
| 0.005926567 |
Beta theo CAPM
# Excess returns
stock_excess_returns <- sweep(stock_returns, 1, rf_daily) # r_i - rf
market_excess_returns <- market_returns - rf_daily # r_m - rf
# Hàm tính beta
calculate_beta <- function(stock_excess_ret, market_excess_ret) {
fit <- lm(as.numeric(stock_excess_ret) ~ as.numeric(market_excess_ret))
unname(coef(fit)[2])
}
betas <- apply(stock_excess_returns, 2, calculate_beta, market_excess_ret = market_excess_returns)
beta_df <- data.frame(Ticker = names(betas), Beta = as.numeric(betas), row.names = NULL)
kable(beta_df, caption = "Hệ số Beta ước tính (CAPM)")
Hệ số Beta ước tính (CAPM)
| HPG |
1.2041160 |
| MWG |
1.1065076 |
| FPT |
0.9289243 |
| VCB |
0.9520191 |
| VNM |
0.6656178 |
TSSL kỳ vọng theo
CAPM
mean_mkt_daily <- mean(as.numeric(market_returns), na.rm = TRUE)
mean_mkt_annual <- mean_mkt_daily * 252
expected_returns_annual <- rf_annual + betas * (mean_mkt_annual - rf_annual)
er_df <- data.frame(
Ticker = names(expected_returns_annual),
`Expected Return (Annual)` = as.numeric(expected_returns_annual)
)
kable(er_df, caption = "TSSL kỳ vọng theo CAPM (annualized)")
TSSL kỳ vọng theo CAPM (annualized)
| HPG |
0.1046185 |
| MWG |
0.0997856 |
| FPT |
0.0909931 |
| VCB |
0.0921366 |
| VNM |
0.0779562 |
Ma trận hiệp phương sai
(annualized)
cov_matrix <- cov(stock_returns, use = "pairwise.complete.obs") * 252
preview_idx <- 1:min(5, ncol(cov_matrix))
kable(round(cov_matrix[preview_idx, preview_idx], 6),
caption = "Một phần ma trận hiệp phương sai (annualized)")
Một phần ma trận hiệp phương sai (annualized)
| HPG |
0.107866 |
0.048645 |
0.039525 |
0.039440 |
0.023099 |
| MWG |
0.048645 |
0.112933 |
0.042372 |
0.032838 |
0.025200 |
| FPT |
0.039525 |
0.042372 |
0.064369 |
0.029260 |
0.023392 |
| VCB |
0.039440 |
0.032838 |
0.029260 |
0.097065 |
0.023125 |
| VNM |
0.023099 |
0.025200 |
0.023392 |
0.023125 |
0.068536 |
Tóm tắt rủi ro &
lợi nhuận
std_annual <- sqrt(diag(cov_matrix)) # độ lệch chuẩn năm
summary_df <- data.frame(
Ticker = names(betas),
`Expected Return (Annual)` = as.numeric(expected_returns_annual[names(betas)]),
`Risk (Stdev Annual)` = as.numeric(std_annual[names(betas)]),
`Beta` = as.numeric(betas),
row.names = NULL,
check.names = FALSE
)
kable(summary_df, digits = 4, caption = "Tóm tắt rủi ro & lợi nhuận theo mã")
Tóm tắt rủi ro & lợi nhuận theo mã
| HPG |
0.1046 |
0.3284 |
1.2041 |
| MWG |
0.0998 |
0.3361 |
1.1065 |
| FPT |
0.0910 |
0.2537 |
0.9289 |
| VCB |
0.0921 |
0.3116 |
0.9520 |
| VNM |
0.0780 |
0.2618 |
0.6656 |
Tối ưu hoá đường biên
hiệu quả (ROI + quadprog, w ≥ 0, ∑w=1)
library(Matrix)
library(ROI)
library(ROI.plugin.quadprog)
# Đầu vào tối ưu hoá (dạng daily)
mu_daily <- as.numeric(expected_returns_annual) / 252
names(mu_daily) <- names(expected_returns_annual)
Sigma_daily <- cov(stock_returns, use = "pairwise.complete.obs")
Sigma_pd <- as.matrix(nearPD(Sigma_daily, corr = FALSE)$mat) # làm dương xác định
# Lưới TSSL mục tiêu
min_ret <- min(mu_daily) + 1e-6
max_ret <- max(mu_daily) - 1e-6
target_returns <- seq(min_ret, max_ret, length.out = 60)
solve_markowitz <- function(target_return, mu, Sigma) {
n <- length(mu)
Q <- 2 * Sigma
L <- rep(0, n)
A <- rbind(rep(1, n), mu) # ∑w=1; mu'w = target_return
dir <- c("==", "==")
rhs <- c(1, target_return)
op <- OP(
objective = Q_objective(Q = Q, L = L),
constraints = L_constraint(L = A, dir = dir, rhs = rhs),
bounds = V_bound(li = 1:n, lb = rep(0, n)) # w >= 0
)
sol <- ROI_solve(op, solver = "quadprog")
w <- sol$solution
names(w) <- names(mu)
risk <- sqrt(sum(w * (Sigma %*% w)))
c(Return = target_return, Risk = risk, w)
}
ports <- lapply(target_returns, function(tr) {
tryCatch(solve_markowitz(tr, mu_daily, Sigma_pd), error = function(e) NULL)
})
frontier_df <- do.call(rbind, ports) %>% as.data.frame()
frontier_df <- frontier_df[is.finite(frontier_df$Return) & is.finite(frontier_df$Risk), , drop = FALSE]
# Annualize & Sharpe
frontier_df$Return_annual <- frontier_df$Return * 252
frontier_df$Risk_annual <- frontier_df$Risk * sqrt(252)
frontier_df$Sharpe_Ratio <- (frontier_df$Return_annual - rf_annual) / pmax(frontier_df$Risk_annual, 1e-12)
# MVP & Max Sharpe
mvp_row <- frontier_df[which.min(frontier_df$Risk_annual), , drop = FALSE]
max_sharpe_row <- frontier_df[which.max(frontier_df$Sharpe_Ratio), , drop = FALSE]
format_portfolio <- function(port_row, caption_text) {
meta <- c("Return","Risk","Return_annual","Risk_annual","Sharpe_Ratio")
tick_cols <- setdiff(colnames(port_row), meta)
w <- as.numeric(port_row[, tick_cols])
names(w) <- tick_cols
w <- w[w > 1e-4]
weights_tbl <- data.frame(`Cổ phiếu` = names(w), `Tỷ trọng` = paste0(round(100*w, 2), "%"))
stats_tbl <- data.frame(
`Thuộc tính` = c("TSSL (năm)", "Rủi ro - ĐLC (năm)", "Sharpe"),
`Giá trị` = c(paste0(round(100*port_row$Return_annual, 2), "%"),
paste0(round(100*port_row$Risk_annual, 2), "%"),
round(port_row$Sharpe_Ratio, 2))
)
print(kable(weights_tbl, caption = caption_text, row.names = FALSE))
print(kable(stats_tbl, caption = "Thông số danh mục", row.names = FALSE))
}
format_portfolio(mvp_row, "Danh mục rủi ro tối thiểu (MVP) - Tỷ trọng")
##
##
## Table: Danh mục rủi ro tối thiểu (MVP) - Tỷ trọng
##
## |Cổ.phiếu |Tỷ.trọng |
## |:--------|:--------|
## |HPG |8.91% |
## |MWG |7.18% |
## |FPT |31.07% |
## |VCB |17.31% |
## |VNM |35.52% |
##
##
## Table: Thông số danh mục
##
## |Thuộc.tính |Giá.trị |
## |:------------------|:-------|
## |TSSL (năm) |8.84% |
## |Rủi ro - ĐLC (năm) |19.96% |
## |Sharpe |0.22 |
format_portfolio(max_sharpe_row, "Danh mục Sharpe tối đa - Tỷ trọng")
##
##
## Table: Danh mục Sharpe tối đa - Tỷ trọng
##
## |Cổ.phiếu |Tỷ.trọng |
## |:--------|:--------|
## |HPG |23.79% |
## |MWG |15.29% |
## |FPT |27.28% |
## |VCB |17.88% |
## |VNM |15.78% |
##
##
## Table: Thông số danh mục
##
## |Thuộc.tính |Giá.trị |
## |:------------------|:-------|
## |TSSL (năm) |9.37% |
## |Rủi ro - ĐLC (năm) |21.17% |
## |Sharpe |0.23 |