Novel Approaches to Identify Clusters Using Independent Components Analysis with Application

Table 4

The ICA-based clusters using ranked approach (case: FastICA).

Cluster no.	Cluster size	Companies
1	39	AGIC, AGTL, ANL, BNWM, CHCC, COTT, DNCC, DWSM, EFUL, ESBL, FABL, FCCL, FCSC, FECTC, FFLM, FHBM, GHNL, HCAR, HICL, HINO, HINOON, INIL, JDWS, JSCL, KASBB, KOHC, LUCK, MARI, MTL, NATF, NBP, PAKRI, PIAA, PNGRS, POL, PRL, SAZEW, SEARL, and SIEM
2	10	ABOT, AICL, CLOV, DAWH, ICI, IDRT, MIRKS, NESTLE, OLPL, and PSMC
3	8	ATLH, ATRL, BWHL, DFML, GASF, MCB, MODAM, and PTEC
4	7	CPL, EFUG, FUDLM, JOVC, NCL, PCAL, and REWM
5	6	BAHL, KOHP, KTML, MACFL, SAIF, and SANSM
6	5	GWLC, FNBM, NICL, PAKD, and QUICE
7	4	CENI, KSBP, PICT, and SIBL
8	4	DYNO, ENGRO, IDYM, and SHFA
9	4	FDIBL, FEM, JOPP, and PAEL
10	3	DGKC, JPGL, and MUREB
11	3	FFBL, GLAXO, and HADC
12	3	HSPI, NIB, and RICL
13	2	ACPL and MEBL
14	2	AGIL and GLPL
15	2	BYCO and DSFL
16	2	CEPB and GTYR
17	2	CHAS and GHGL
18	2	FPJM and PMI
19	2	HAL and SEPCO
20	2	HUBC and PKGS
21	2	KESC and PSO
22	2	MLCF and SCM