Introduction

In this experiment, I will apply Apriori and FP-Growth method for association rule mining.

Exploratory analysis

Each row is a plant.
First column is Latin name (species or genus), the others are states distribution.
Binary value (y/n). y means exist in this state, n means doesn't exist in this state.
Save data in csv format: plants.csv

Generated sets of large itemsets:

Size of set of large itemsets L(1): 49

Size of set of large itemsets L(2): 167

Size of set of large itemsets L(3): 120

Size of set of large itemsets L(4): 25

Size of set of large itemsets L(5): 2

Best rules found:

ct=y ma=y nj=y 3562 ==> ny=y 3524 conf:(0.99)
tn=y md=y nc=y 3531 ==> va=y 3489 conf:(0.99)
ct=y pe=y nj=y 3618 ==> ny=y 3571 conf:(0.99)
ga=y va=y al=y sc=y 3579 ==> nc=y 3529 conf:(0.99)
ga=y va=y sc=y 3882 ==> nc=y 3825 conf:(0.99)
ms=y al=y nc=y sc=y 3572 ==> ga=y 3519 conf:(0.99)
nj=y ny=y oh=y 3556 ==> pe=y 3502 conf:(0.98)
ct=y pe=y ma=y 3784 ==> ny=y 3726 conf:(0.98)
pe=y ma=y nj=y 3699 ==> ny=y 3640 conf:(0.98)
ga=y md=y nc=y 3551 ==> va=y 3484 conf:(0.98)

FPGrowth found 107 rules (displaying top 10)

[ma=y, nj=y, ct=y]: 3562 ==> [ny=y]: 3524 conf:(0.99) lift:(5.96) lev:(0.08) conv:(76.17)
[nc=y, md=y, tn=y]: 3531 ==> [va=y]: 3489 conf:(0.99) lift:(6.1) lev:(0.08) conv:(68.81)
[pe=y, nj=y, ct=y]: 3618 ==> [ny=y]: 3571 conf:(0.99) lift:(5.95) lev:(0.09) conv:(62.86)
[ga=y, al=y, va=y, sc=y]: 3579 ==> [nc=y]: 3529 conf:(0.99) lift:(5.79) lev:(0.08) conv:(58.22)
[ga=y, va=y, sc=y]: 3882 ==> [nc=y]: 3825 conf:(0.99) lift:(5.78) lev:(0.09) conv:(55.53)
[nc=y, al=y, sc=y, ms=y]: 3572 ==> [ga=y]: 3519 conf:(0.99) lift:(5.77) lev:(0.08) conv:(54.85)
[ny=y, nj=y, oh=y]: 3556 ==> [pe=y]: 3502 conf:(0.98) lift:(6.21) lev:(0.08) conv:(54.4)
[pe=y, ma=y, ct=y]: 3784 ==> [ny=y]: 3726 conf:(0.98) lift:(5.93) lev:(0.09) conv:(53.49)
[pe=y, ma=y, nj=y]: 3699 ==> [ny=y]: 3640 conf:(0.98) lift:(5.93) lev:(0.09) conv:(51.42)
[ga=y, nc=y, md=y]: 3551 ==> [va=y]: 3484 conf:(0.98) lift:(6.05) lev:(0.08) conv:(43.76)