Hi,

Please find the answer below.

How can I specify label of new data according to the weather.nominal.arff data in Weka?

I open the weather.nominal.arff data and classify it with J48, but I don't know how can I specify label of new data like this:

@relation NewData

@attribute outlook {sunny,overcast,rainy}

@attribute temperature {hot,mild,cool}

@attribute humidity {high,normal}

@attribute windy {TRUE,FALSE}

@attribute play {yes,no}

@data

sunny,hot,high,FALSE

sunny,hot,high,TRUE

overcast,hot,high,FALSE

rainy,mild,high,FALSE

rainy,cool,normal,FALSE

rainy,cool,normal,TRUE

overcast,cool,normal,TRUE

Answer

In this example data, any instance must have one class label (yes/no). In this case, the newly added instance must be assigned to one of these two labels. Here is an example with label 'yes' added to your 1st instances of your data:

sunny,hot,high,FALSE, yes

HTH.

Dr. Samer Sarsam

I suggest you to follow https://www.researchgate.net/post/How_to_classify_a_new_unlabel_test_data_in_weka

See its tutorial too : https://www.youtube.com/watch?v=uiDFa7iY9yo

1. I convert coloured image into gray scale

2. Using this code: detectSURFFeatures(imagesGray);, I extract the key interest point

3.clustered the extraction point using kmeans

Here is the link which I referred:

https://www.mathworks.com/help/vision/ref/detectsurffeatures.html

Samer Sarsam

Hi Samer,

I want to specify label of new data with WEKA. I used that simple example to show how can I do it with WEKA.

In other words, I don't know how I should input unlabeled data in the WEKA and specify labels (with WEKA).

I will be grateful if you explain step by step.

Thanks

Muhammad Ali

Hi Muhammad,

I saw it. Thanks. But I don't know what I should do, like how input unlabeled data to WEKA and classify it based on some WEKA algorithms (with WEKA).

Bahman Mirzakhani

If the new example has a label (e.g., yes/no), then you can define it as the below example where the instance here has a 'yes' label:

sunny,hot,high,FALSE, yes

If the new example does not have a label, then we use the '?' such to indicate a missing label of the instance. Here is an example instance with the missing label:

sunny,hot,high,FALSE, ?

Note that the missing value has to be predicted by the machine learning algorithm after learning your training set properly from full labeled data.

Dr. Samer Sarsam

Samer Sarsam

Thank you. I put question marks in the last column for unlabeled data, but how can I classify them? Assume I have a learned algorithm.

My main problem is prediction of label of unlabeled data by WEKA! Can WEKA do?

Samer Sarsam

Hi again

May I ask why was your last answer deleted?

It was helpful for me.