README.md

In celehs/IFPCA: Intensity Functional Principal Component Analysis (IFPCA)

Intensity Functional Principal Component Analysis

Overview

This package re-implements the first step of the {MASTA} package to extract features from longitudinal encounter records. Compared to {MASTA}, the input data of {IFPCA} is more compact and memory efficent. Click HERE to view input data structure.

Installation

Install development version from GitHub.

# install.packages("remotes")
remotes::install_github("celehs/IFPCA")

Load the package into R.

library(IFPCA)

Data Example

Data Preparation

library(data.table)

i <- 3 # 1, 2, 3 
# time for each code (training + validation)
url <- "https://raw.githubusercontent.com/celehs/IFPCA/master/data-raw/"
time_code <- fread(paste0(url, "time_code", i, ".csv"))
time <- time_code$month
names(time) <- time_code$id
# follow up time for training and validation sets
follow_up_train <- fread(paste0(url, "follow_up_train.csv"))
follow_up_valid <- fread(paste0(url, "follow_up_valid.csv"))
fu_train <- follow_up_train$fu_time
fu_valid <- follow_up_valid$fu_time
names(fu_train) <- follow_up_train$id
names(fu_valid) <- follow_up_valid$id
str(fu_train)

##  Named num [1:20600] 49.4 13.9 12.6 14.9 80.7 ...
##  - attr(*, "names")= chr [1:20600] "1" "2" "3" "4" ...

str(fu_valid)

##  Named num [1:500] 71.7 70.4 14.9 33.8 98.6 ...
##  - attr(*, "names")= chr [1:500] "90001" "90002" "90003" "90004" ...

Feature Extraction

system.time(ans <- ifpca(time, fu_train, fu_valid))

##    user  system elapsed 
##   8.275   0.332   8.614

data.table(ans$TrainFt) # Extracted Features (Training) 

##         1stCode        Pk       ChP    1stScore      logN
##     1: 49.41273 49.412731 49.412731 -1.77557902 0.0000000
##     2: 13.93018 13.930185 13.930185 -1.77557902 0.0000000
##     3: 12.55031 12.550308 12.550308 -1.77557902 0.0000000
##     4: 14.85010 14.850103 14.850103 -1.77557902 0.0000000
##     5: 80.65708 80.657084 80.657084 -1.77557902 0.0000000
##    ---                                                   
## 20596:  2.00000  3.750801  3.595729 -0.05899512 1.3862944
## 20597: 14.00000 14.359918 14.140123  0.71721892 0.6931472
## 20598:  8.00000  7.638604  7.434908 -1.32323471 1.0986123
## 20599: 13.00000 13.059548 12.711294 -1.36421318 1.0986123
## 20600: 23.00000 24.274004 23.841643  3.02440941 1.3862944

data.table(ans$ValidFt) # Extracted Features (Validation)

##       1stCode       Pk      ChP  1stScore      logN
##   1: 71.72074 71.72074 71.72074 -1.775579 0.0000000
##   2: 70.37372 70.37372 70.37372 -1.775579 0.0000000
##   3: 14.94867 14.01437 13.64066  3.038376 0.6931472
##   4: 33.80698 33.80698 33.80698 -1.775579 0.0000000
##   5: 98.62834 98.62834 98.62834 -1.775579 0.0000000
##  ---                                               
## 496: 35.58111 35.58111 35.58111 -1.775579 0.0000000
## 497: 27.36756 27.36756 27.36756 -1.775579 0.0000000
## 498: 24.60780 24.60780 24.60780 -1.775579 0.0000000
## 499: 27.17043 27.17043 27.17043 -1.775579 0.0000000
## 500: 29.30595 29.30595 29.30595 -1.775579 0.0000000

References

• Wu, S., Müller, H., Zhang, Z. (2013). Functional Data Analysis for Point Processes with Rare Events. Statistica Sinica, 23:1-23. https://doi.org/10.5705/ss.2010.162

• Liang, L., Uno, H., Ma, Y., Cai, T. Robust Approach to Event Time Annotation Using Longitudinal Medical Encounters. Working Paper.

celehs/IFPCA documentation built on Dec. 17, 2020, 10:21 p.m.