How can we select specific attributes using WEKA API?

More Noura H. Al Nuaimi's questions See All

Publisher Accept Survey Paper - Journal Suggestion?

I have a survey paper on feature selection and big data. I need a suggestion for a journal that accepts the survey. I am looking for a good impact factor and review period of around one month.

06 July 2018 7,364 0 View

Journals that accept survey papers?

Any suggestion for journals that accept survey papers in data mining/big data? Short reviewing period max 2 months?

10 November 2017 1,844 0 View

Get Linkedhashmap item index?

Is there a way to get the index of Linkedhashmap key or value?

07 August 2016 5,030 0 View

K-greedy algorithm JAVA?

Any recommendation for k-greedy algorithm implemented in JAVA for data mining?

04 May 2016 5,652 4 View

Feature Selection for big data?

I am doing a research on feature selection for extra-high dimensional data (big data). Does any one have good advice about finding the redundant features? Please don't mention PCA, decision trees...

04 May 2016 9,164 7 View

Top Ten Learning Algorithm?

01 January 1970 2,809 34 View

How to compare two groups with only two measurements?

Hiiiii everyone! I have an enquiry on statistical analysis. I was looking for many forum and it's still cannot solve my problem. I want to compare means of two groups of data but only with two...

03 March 2021 8,796 3 View

What's the best way to measure growth rates in House sparrow chicks from day 2 to day 10?

What's the best way to measure growth rates in House sparrow chicks from day 2 to day 10? Since, the growth curve from day 2 to 10 won't be like the "Logistic curve" it might not follow logistic...

03 March 2021 1,401 3 View

How to increase a model accuracy ?

dear community, my model is based feature extraction from non stationary signals using discrete Wavelet Transform and then using statistical features then machine learning classifiers in order to...

03 March 2021 6,994 5 View

1. What is the impact of having different scales in a survey? and how can we solve this problem before and after data collection (Literature)?

1. What is the impact of having different scales in a survey? and how can we solve this problem before and after data collection (Literature-based reflection)? Thank you for your time and for...

03 March 2021 2,870 3 View

What degree of missing data in a particular case/participant results in Expectation-Maximisation no longer working?

I'm dealing with a mediation model and am using the PROCESS module in SPSS. Due to SPSS and PROCESS being limited in the imputation methods - being unable to handle multiple imputation - the other...

02 March 2021 4,362 1 View

Can NFL theorem be valid in infinite search space in ML?

NFL theorem is valid for algorithms training in fixed training set. However, the general characteristic of algorithms in expanded or open dataset has not been proved yet. Could you show your...

01 March 2021 1,189 3 View

Does anyone know how to plot my station to a section plot on Ocean Data View (ODV)?

I made a vertical section plot with WOA .nc file on ODV, and now I want to plot my stations (csv. file) in it. Does anyone know how to import my points to the section? They are two different...

01 March 2021 3,610 1 View

How to address this issue in the analysis of Real-Time PCR results?

To dear Researchers, I was analyzing a series of concentration for estimation of Real-Time PCR efficiency. The concentration was 1:10. I used MS-excel to evaluate Slope. The result of slope was -8...

01 March 2021 8,683 4 View

How to arrange coordinates and extract climate dataset out of MPI-ESM-MR netcdf files?

Have a nice day everyone, I'm stuck in extracting datasets out of MPI-ESM-MR netcdf files because the latitude ranges from -3*10^6 to 3*10^6 and longitude ranges -5,9*10^6 to 5,9*10^6. Plus, the...

28 February 2021 912 3 View

Is there any Machine Learning/AI algorithm that lets you predict with incomplete features without missing value imputation process?

For instance, the model is trained with A,B,C,D,E features to predict F target. Then, I want to make prediction on a new data, but with only A,C,D,E is known without doing missing value imputation...

28 February 2021 1,556 6 View

Samer Sarsam

Hi Noura,

Why don't you use WEKA GUI to perform this task?

However, you could use the "ArffSaver" class through the API, in order to solve your issue.

HTH.

Samer

Noura H. Al Nuaimi

Thanks Samer. I am using WEKA for handling arff files and do my own processing on the data with Java.

I have a look at ArffSaver, it save the Instances object. I need a way to select specific attributes from the Instances object and save them with the class.

Tunc Guven Kaya

You might use Attribute Selection class.

Thanks Tunc Guven Kaya, but Attribute Selection class use specific algorithm to select best attributes.

For my case, I want to select like 1st, 2nd and last attributes + class and save them to new Instances object.

Noura,

Before going through the code, in order to understand your case correctly, I attached here two files represent the loaded iris dataset. First one shows the whole 5 attributes (including the class attribute), while the second one shows only 3 attributes (with the class) after removing first two attributes from the data. Now, are you asking about the way of saving those 3 attributes using API?

NB: AGAIN, I know that you need to work from the API, but my aim is understanding your situation from the GUI in order to provide a proper help.

Sorry, in Weka under the preprocess tab, you can choose filters. Under "unsupervised-->attribute" part, you can see some options like remove, removeByName etc.

Samer Sarsam, I am familiar with the WEKA GUI. But, I need to do it on WEKA API in order to extend extra functionality.

Tunc Guven Kaya, Yes you are right. I only found this way on WEKA API that allow removing un wanted attributes from the Instances Object. But, I am looking for another way to do.

I will give an example. Let say I have a dataset with 50 attributes. I need to select only 10 and save them to new ARFF file.

The limitation with WEKA GUI, it allow this by deleting the unwanted attributes.

I am wondering if there is an alternative to do it with WEKA API?

Noura H. Al Nuaimi,

Well, seems like you didn't read my previous post correctly, where I clearly mentioned that my major aim was *only* understanding your issue from the GUI, in order to help you with the API side (not asking you using the GUI).

Anyway, back to your question. Here are two solutions:

First solution: adjusting the ARFF relation, but it would require more code and would be more likely to yield bugs.

Second solution: using the "inverse" mode of the Remove filter. Something like this:

import weka.core.Instances;

import weka.filters.Filter;

import weka.filters.unsupervised.attribute.Remove;

...

Instances data = ... ; // your data with last attr as class

Remove remove = new Remove();

remove.setAttributeIndices("1,8,last");

remove.setInvertSelection(true);

remove.setInputFormat(data); // init filter

Instances filtered = Filter.useFilter(data, remove); // apply filter

Thanks Samer for your support. I got your idea :)

Hi,

This won't be a direct answer to your question but as a Weka laborer for years I find it much easier to perform some certain tasks on other platforms like NotePad++, LibreOffice etc.

For example, if you open the data part of the arff file in libreoffice or excel, you can delete the columns you do not need and that gives you the solution for this specific problem.

Moreover, you can perform column operations to create new features or manipulate the existing ones. Let's assume that you have 3 numerical values as your features: f1, f2, f3. You can perform multiplication like f1*f2 on the next column. You can even abstract some higher-level features like NDVI value for satellite imagery by doing more complicated calculations.

By using NotePad++, you can delete some attributes line by line by using regular expressions. Or you can merge some labels with "find & replace" command.

By using different tools, you can also visualize your data or do normality tests, t-tests etc.

In a nutshell, data part of the arff is just csv format and you can copy that part into some other software where you feel more talented and competent and perform whatever you need.

Hope this helps.