Data Files - CS5483

# initialization
import os

if not os.getenv(
    "NBGRADER_EXECUTION"
):  # Skip the code or auto-grading may take too long to complete.
    %reload_ext jupyter_ai

Load data into Weka Explorer Interface¶

How to start Weka in JupyterHub?

Open the Launcher (File → New Launcher)
Start a Desktop from the Launcher.
Start a Terminal from the menu on the top left.
Run the command weka and click the Explorer button.
Load data from the folder /data/ under the linux root directory.

%%ai
How is Weka implemented and what is its main advantage over other data mining 
tools?

Other ways to run Weka

To run Weka locally on your computer:
1. Download and install Weka from here, and
2. obtain the data files from
  1. the subfolder data of your Weka installation path, or
  2. download from here.
For the computers in CSC teaching studios, you can start Weka as follows:
1. Click the shortcut Work Desk from desktop.
2. Click the link Weka 3.8.x for CS Department.
3. Load the dataset from
  C:\Program Files\Weka-3-8\data\.
For the computers in CS labs, you can start Weka as follows:
1. Execute G:\weka\3.8\run.bat,
2. Click the Explorer button, and
3. Load the dataset from
  C:\temp\Weka-3.8\data or G:\weka\3.8\files\data.

Use Weka to do [Witten11] Exercises 17.1.1 and 17.1.2..

YOUR ANSWER HERE

Create an ARFF file¶

Exercise 3

Create an ARFF file named AND.arff in the current directory for the AND gate $Y=X_1\cdot X_2$ . Use 0 and 1 to represent False and True respectively.

Tip

To generate the file from this notebook directly:

Copy the solution template below to the following solution cell.
Replace each underscore _ by an appropriate value.
Execute the code cell.

content = '''
@RELATION AND
@ATTRIBUTE X1 {0, 1}
@ATTRIBUTE X2 {_, _}
@ATTRIBUTE Y {_, _}
@DATA
_, _, _
_, _, _
_, _, _
_, _, _
'''

Alternatively, you may use other editors in JupyterHub to create AND.arff in the current folder.

Hint

See the documentation for the ARFF file format, or take a look at some of the ARFF files in the WEKA data folder as examples.

# YOUR CODE HERE
raise NotImplementedError

# write the content of text to the file
try: content
except NameError: 
    print("AND.arff not generated because `content` is undefined.")
else:
    filename = 'AND.arff'
    with open(filename,'w') as f:
        f.write(content)
    print("AND.arff generated.")

Run the following test cell to see if your file is a valid ARFF file. You may also download and load the ARFF file into WEKA to see if there is any syntax error.

# test
print('Content of AND.arff:')
with open(filename) as f:
    print(f.read())

from scipy.io import arff
import pandas as pd

d = arff.loadarff(filename)
df = pd.DataFrame(d[0]).astype(int)
df.head()

%%ai
How is ARFF compared to CSV is Weka implemented and why is one former better or
more popular than the other?