Page 253 - Ai_C10_Flipbook
P. 253
The sample data taken for our steps can be download from link.
https://orangewebsupport.co.in/AI/IMDB_Dataset_ODM.xlsx
OR
http://bit.ly/4jLNdhG
Steps to teach AI model to predict the sentiments at different stages of Project life cycle using the Orange Data
Mining tool are as follows:
Stage 1 Problem Scoping
From the IMDb dataset, predict for each review in the corpus. Predict the following sentiments from the dataset:
Sentiment Meaning
Positive Score Positive Review
Negative Score Negative Review
Neutral Score Neutral Review
Compound Score Total Sentiment of the review i.e. -1 means most negative, +1 means most positive
Stage 2 Data Acquisition
Follow the following steps in the process of data acquisition:
Step 1 Open the Orange Data Mining application and click on the New option. It opens the new canvas.
Step 2 Click on File widget to place it on the Canvas.
Step 3 Click on Corpus to place it on the Canvas and load the dataset from the computer in the xls format to
the corpus. Note, Text Mining can be added by selecting Text option from Add-ins from Options tab.
Step 4 Connect the File to Corpus by dragging the output from file to the input of Corpus.
Step 5 Insert the Data Table just next to Corpus and connect the Corpus to Data Table by dragging the
output from Corpus to the input of Data Table.
Step 6 Double-click on Data Table widget to see how data appears now.
Step 7 Insert the Corpus Viewer just next to Corpus and connect the Corpus to Corpus Viewer by dragging
the output from Corpus to the input of Corpus Viewer.
Step 8 Double-click on the Corpus Viewer widget to see the data.
Stage 3 Data Exploration
Follow the following steps in the process of data exploration:
Step 9 Double-click on Data Table widget to see missing data.
Step 10 Insert the Preprocess Text and connect Corpus to Preprocess Text by dragging the output from
Corpus to the input of Preprocess Text.
Step 11 Double-click on PreProcess Text to open the Properties Window. We can see that the data is cleaned
automatically.
Stage 4 Modelling
To model the stated dataset follow the given step:
Step 12 Insert the Sentiment Analysis and connect the PreProcess Text to Sentiment Analysis by dragging the
output from PreProcess Text to the input of Sentiment Analysis.
Natural Language Processing 251

