Can't load package FSelector for hill.climbing.search

Question

I want to use the hill.climbing.search on my data to detect the most significant deviation in my data. Therefore, I wanted to install the package FSelector which works, but when loading the package I receive this error message:

> library(FSelector)
Error: package or namespace load failed for ‘FSelector’:
 .onLoad in loadNamespace() for 'rJava' failed, Details:
  Call: fun(libname, pkgname)
  Error: JAVA_HOME cannot be determined from the Registry

In addition, I read the introduction to hill.climbing.search from the pdf from CRAN for FSelector package, but I am confused what in my data the parameters attributes and eval.fun are.

Here is my data:

> df1
      date               count
1  2012-07-01           2.327684
2  2012-08-01           1.351609
3  2012-09-01           1.532830
4  2012-10-01           2.135562
5  2012-11-01           2.847107
6  2012-12-01           3.585103
7  2013-01-01           4.375418
8  2013-02-01           4.138450
9  2013-03-01           5.929983
10 2013-04-01           5.178621
11 2013-05-01           6.756757
12 2013-06-01           4.749200
13 2013-07-01           2.208296
14 2013-08-01           2.244003
15 2013-09-01           2.016025
16 2013-10-01           1.698015
17 2013-11-01           1.459854
18 2013-12-01           3.350970