Following on from the Advanced Data Mining with WEKA .. lesson 3.2 .. I am
trying to install the weka support for R under linux mint 18 ... to date I
have been happily using weka and moa ..
in this OS.
I have the RConsole running but when I try to plot the iris data as per the
video lesson I get the following error ..
"R does not seem to be available." ... my shell script for running weka is
java -Xmx2000M -jar weka/weka.jar
and I can run R from the command line in a bash shell.
I have attached a listing of the command
ls -alR /usr/lib/R as path_contents.txt.
~/R contains the directory "x86_64-pc-linux-gnu-library"
Many thanks for any advice ... maybe I have misinterpreted R_HOME?
(the R binary is also in /usr/bin)
I am trying to use RConsole with Weka.
Please provide an example for doing k-means clustering of the loaded
dataset, on the RConsole of weka on the Explorer.
This will be very helpful to understand the GUI.
Thank you very much.
View this message in context: http://weka.8497.n7.nabble.com/Example-for-R-PLugin-tp30563.html
Sent from the WEKA mailing list archive at Nabble.com.
I am running J48 on weather.nominal file with default parameters
I have attached the screen shot of *Visualize Classifier Errors* window
and clicking on one of the squares.
The instance numbers given in the panel do not correspond to the instance
numbers viewed in the edit window.
Why are they different and what am missing (if I am missing something)?
Instance #6 in the panel corresponds to instance #4 in the viewer
Instance #8 in the panel corresponds to instance #12 in the viewer
I'm now trying to connect Weka to Mysql database. But after I select data in
SQl-viewer and press "Ok" I get this error:
> couldn't read from database unknown data type: INT. Add Entry in
> If the type contains blanks, either escape them with a backslash or
> use underscores instead of blanks.
> I read several treads about this error but still cannt find out what
> is wrong. I would be great if someone take a look at my
> DatabaseUtils.props and give me a hint what is wrong
Here is my DatabaseUtils.props:
# General information on database access can be found here:
# Version: $Revision: 5836 $
# The comma-separated list of jdbc drivers to use
# The url to the experiment database
#the method that is used to retrieve values from the db
# (java datatype + RecordSet.<method>) string, getString() = 0; --> nominal
boolean, getBoolean() = 1; --> nominal
double, getDouble() = 2; --> numeric byte, getByte() = 3; --> numeric
short, getByte()= 4; --> numeric
int, getInteger() = 5; --> numeric INT, getInteger() = 5; INT(11),
getInteger() = 5; -->numeric INT., getInteger() = 5; -->numeric long,
getLong() = 6; --> numeric float, getFloat() = 7; --> numeric
date, getDate() = 8; --> date text, getString() = 9; --> string
time, getTime() = 10; --> date
# the original conversion: <column type>=<conversion>
#mappings for table creation
# All the reserved keywords for this database
# The character to append to attribute names to avoid exceptions due to
# clashes between keywords and attribute names
#flags for loading and saving instances using DatabaseLoader/Saver
I have two numeric attributed (no class), and I want to use Apriori. I
applied unsupervised discretize filter in the preprocess panel, so it
convert only first attribute to nominal but not the second one because it
considered it as a class.
My question, what is the parameter of the unsupervised descrerize that
should I configure to allow me discretize both attributes?
P.S.: I prefer discretize in my case instead of using the filter
Thanks in advance.
My question is as follows:
I would like to use the discretization process in my database using Weka.
In filter has the option to discretize the supervised and unsupervised
approach. I will use the PCA method (Principal Component Analysis) later.
So what would be the option to discretization in which approach and why?
I am using weka3.5.6 developer version.When i run weka with console,it says some error messages "trying to add JDBC driver:rmijdbc.Rjdriver-error,not in CLASSPATH? etc..i can't get what it is..
As I mentioned before, I am new to Weka, and so I am using my dataset to
perform comparisons among different classifier types (rules verses trees
verses SVMs verses Bayesian, ETC) One of the classifiers I'm testing is
I am hoping that someone can help me figure out the best way to handle
I ran the JRip classifier under the WeightedInstancesHandlerWrapper
metaclassifier on a training file with 10 fold CV and two optimization runs
as any other number seemed to degrade performance.
Pruning is set to on by default.
I ended up with 18 rules, but there is a confusing behavior in the
outputted rules. Can someone tell me if this is normal?
(fifth <= 5.874) and (fifth <= 4.655) and (first <= 3235) => class=a
Please notice "fifth". Is that second "<" sign meant to be a ">" sign? If
not, why doesn't the rule simply use one value or the other?
(third >= 0.6914) and (fourth >= 14558) and (first >= 144) and (third >=
0.8019) and (fifth <= 15.23) => class=a (305.0/0.0)
Same with "third" here. Is the second ">" meant to be a "<" ??
=> class=b (5138.0/40.0)
I assume this means that anything which doesn't match the rules for "a"
goes into "b" ?
If the rules are correct as they stand, then do I just manually "fix" them
to get rid of the extraneous comparison? If so, how do I know which one is
the extra one?
Thanking you in advance,