clusEvol: A Procedure for Cluster Evolution Analytics.

About

The objective of CEA is to analyze the evolution of an object and its neighbours over time. Neighbours are identified by clustering algorithms, CEA leverages the temporal component of panel data and it is based on combining two techniques that are usually not related: leave-one-out and plug-in principle. This allows us to use exploratory what if questions in the sense that the present information of an object is plugged in a dataset in a previous timepoint so that we can explore its evolution (and that of its neighbors) up to the present.

CEA: Cluster Evolution Analytics

Installation Instructions

Option 1: CRAN

Our cross-platform version can be found at CRAN: clusEvol package. You can install it with:

install.packages("GeoModels")

Option 2: Developer

For clusEvol installation you will need to previously install devtools package if you do not have it installed yet:

install.packages("devtools")

library(devtools)

devtools lets you install packages from github since they need to be installed from source code.

Description

clusEvol is a function that allows us to use exploratory what if questions in the sense that the present information of an object is plugged-in a dataset in a previous time frame so that we can explore its evolution (and of its neighbours) to the present.

clusEvol can be synthetized has the following steps

• Identify clusters to which sel.obj is similar in time.base .

• The data of sel.obj in time.base is plugged-in in each time period.

• Clusters are generated in each time period with data from sel.obj in time.base.

Example

library(clusEvol)
data(actpas)

solclusEvol <- clusEvol(x=actpas,objects="razon_social", 
time = "fecha",target.vars = c("montoAct","operAct"),
                        time.base=max(actpas$fecha),
                        sel.obj="BANCO SOLIDARIO S.A.",init = min(actpas$fecha),
                        logscale = TRUE,ng = 5,clm = "pam")
print(solclusEvol)

## 
## ##################################################################
## clusEvol: Cluster Evolution Analytics
## 
## 
## Number of neighbours  BANCO SOLIDARIO S.A. is a group member:  
## 2021-06 2021-07 2021-08 2021-09 2021-10 2021-11 2021-12 2022-01 2022-02 2022-03 
##      11      12      12      29      30      28      11      24      13      13 
## 2022-04 2022-05 2022-06 2022-07 2022-08 2022-09 2022-10 2022-11 2022-12 2023-01 
##      25      12      21      14      10      11      27      44      12      11 
## 2023-02 2023-03 2023-04 2023-05 
##      39      12      11      10 
## 
## 
## Cluster that  BANCO SOLIDARIO S.A. belongs to:  
## 2021-06 2021-07 2021-08 2021-09 2021-10 2021-11 2021-12 2022-01 2022-02 2022-03 
##       5       4       5       3       3       3       4       3       4       5 
## 2022-04 2022-05 2022-06 2022-07 2022-08 2022-09 2022-10 2022-11 2022-12 2023-01 
##       3       4       3       4       4       5       3       3       3       4 
## 2023-02 2023-03 2023-04 2023-05 
##       3       4       4       4 
## 
## Clusters in time:
##          
##              1    2    3    4    5
##   2021-06   64   81   41   90   12
##   2021-07   53   93   83   13   68
##   2021-08   53   88   75   92   13
##   2021-09   55   76   30   86   84
##   2021-10   53   87   31   82   78
##   2021-11   52   82   29   59  110
##   2021-12   47  106   79   12   87
##   2022-01   43   92   25   53  115
##   2022-02   46   96  105   14   73
##   2022-03   43   92   93   90   14
##   2022-04   42   81   26  103   80
##   2022-05   47   92   83   13   96
##   2022-06   51   93   22  100   66
##   2022-07   45   85   86   15  101
##   2022-08   43  102   86   11   93
##   2022-09   44  105   86   91   12
##   2022-10   48  102   28   87   76
##   2022-11   97   91   45   73   40
##   2022-12  101   41   13   77  110
##   2023-01   41   94  110   12   83
##   2023-02   79   87   40   99   44
##   2023-03   54  102  102   13   77
##   2023-04   47  114  106   12   68
##   2023-05   48  113  101   11   76
## attr(,"class")
## [1]  table

Note that the example uses actpas dataset cointained in the package. actpas is Ecuador’s amount of Assets and Liabilities Operations of the National Financial System, openly available at: https://contenido.bce.fin.ec/home1/economia/tasas/IndiceSFN.htm

The print method displays principal aspects of CEA. In this case, BANCO SOLIDARIO S.A. is selected to be analyzed in monthly data from 2021-06 to 2023-05:

• Number of neighbours BANCO SOLIDARIO S.A. is a group member.
• Cluster that BANCO SOLIDARIO S.A. belongs to.
• Clusters in time.

clusEvol function outputs a clusEvol class object that has a plot method associated (?plot.clusEvol for detailed parameter information):

plot(solclusEvol,2)

Plot shows the neighbours of the selected object over time.

Package Citation

Once you have installed clusEvol, you can have a BibTex citation with citation("clusEvol") and get:

To cite package ‘clusEvol’ in publications use:

  Morales-Oñate V, Morales-Oñate B (2024). _clusEvol: A Procedure for Cluster Evolution Analytics_.
  R package version 1.0.0, <https://CRAN.R-project.org/package=clusEvol>.

A BibTeX entry for LaTeX users is

  @Manual{,
    title = {clusEvol: A Procedure for Cluster Evolution Analytics},
    author = {Víctor Morales-Oñate and Bolívar Morales-Oñate},
    year = {2024},
    note = {R package version 1.0.0},
    url = {https://CRAN.R-project.org/package=clusEvol},
  }