Gedevan-Aleksizde

require(KFAS) # 1.2.9
require(dplyr)
require(tidyr)
require(ggplot2)

# --- ARIMA(2, 1) with linear trend ---
# generate a dataset
set.seed(42)
t <- 100
y <- arima.sim(n = t, model = list(ar=c(.3, -0.1),  ma=.2), sd=.1) + seq(from=1, to=10, length.out = t)

Gedevan-Aleksizde

require(bsts) # 0.7.1

data(iclaims)     # bring the initial.claims data into scope

# --- model 1 ----
ss <- AddLocalLinearTrend(list(), initial.claims$iclaimsNSA)
ss <- AddSeasonal(ss, initial.claims$iclaimsNSA, nseasons = 52)
model1 <- bsts(initial.claims$iclaimsNSA,
               state.specification = ss,
               niter = 1000)

Gedevan-Aleksizde

source("common.R", encoding = "utf-8")

df$RP <- calc_RP(df$RP095, df$AP, .95)
z <- calc_Z(RP = df$RP, p = df$AP)
df <- mutate(df, z1=z$z1, z2=z$z2)
df <- mutate(df, z1E=z1*end, z2E=z2*end)

ss <- AddLocalLevel(list(), y = df$logPI) # c
ss <- AddAr(ss, lags=2, y = df$logPI) # AR(2)
# time-varying regression

Gedevan-Aleksizde

source("common.R", encoding = "utf-8")

# ----- KFAS ------
df$RP <- calc_RP(df$RP095, df$AP, .95)
z <- calc_Z(RP = df$RP, p = df$AP)
df <- mutate(df, z1=z$z1, z2=z$z2)
df <- mutate(df, z1E=z1*end, z2E=z2*end)

# specify model
model3KFAS <- SSModel(logPI ~ SSMtrend(1, Q=NA) +

Gedevan-Aleksizde

source("common.R", encoding = "utf-8")

# ---- dlm ------

# model 3
res <- data.frame()
for( a in seq(from=.1, to=.95, by=.05) ){
  RP <- calc_RP(RP = df$RP095, p = df$AP, a = a)
  z <- calc_Z(RP = RP, p = df$AP)
  z1 <- z$z1

Gedevan-Aleksizde

# ------ common part ----

require(ggplot2)
require(dplyr)
require(tidyr)
require(dlm) # 1.1-4
require(KFAS) # 1.2.9
require(bsts) # 0.7.1

# calculate the reference price

Gedevan-Aleksizde

require(mlogit) # 0.2-4

as.mldata <- function(data){
  # convert HC dataset
  # The alternatives are
  # 1. Gas central heat with cooling (gcc)
  # 2. Electric central resistence heat with cooling (ecc)
  # 3. Electric room resistence heat with cooling (erc)
  # 4. Electric heat pump, which provides cooling also (hpc)
  # 5. Gas central heat without cooling (gc)

Gedevan-Aleksizde

# multivariate normal Kalman filter
require(dplyr)
require(tidyr)
require(ggplot2)
require(animation)

# ARIMA(1,1) + 線形トレンド の乱数生成
N      <- 50
phi1   <- .5
theta1 <- .2

Gedevan-Aleksizde

library(dplyr)
library(tidyr)
library(ggplot2)
library(rstan)
library(loo)
library(ggmcmc)

# read datasets
df <- read.csv("rfm.csv", stringsAsFactors = F)
colnames(df)[1] <- "ID"

Gedevan-Aleksizde

/* -----------------------------------------------------------------------------
Estimate latent variable hierarchical Bayes RFM model 

stan ver. 2.14

CREATED: by ill-identified, 18/APR/2016 
REVISED: 25/FEB/2017, REVISED
------------------------------------------------------------------------------- */

data {