Scribd Logo
Upload

Welcome to Scribd!

  • 114 views

    CT107-3-3-TXSA - Group Assignment

    Uploaded by

    Sharveen Veen
    This document outlines a group assignment for a text analytics course. Students will form groups of 3-4 members to complete the following tasks: 1) Build unigram and bigram language models on a provided text corpus to calculate smoothed and unsmoothed probabilities. 2) Calculate sentence probabilities using the language models and compare which model is more suitable. 3) Perform sentiment analysis on music reviews using rule-based and supervised techniques with NLTK and SAS Text Miner, reporting accuracy, precision, recall, and F1 scores. The deliverables are a 2000 word report, Python code, and uploading the work through the LMS. Academic integrity policies are also outlined.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd

    CT107-3-3-TXSA - Group Assignment

    Uploaded by

    Sharveen Veen
    114 views4 pages
    This document outlines a group assignment for a text analytics course. Students will form groups of 3-4 members to complete the following tasks: 1) Build unigram and bigram language models on a provided text corpus to calculate smoothed and unsmoothed probabilities. 2) Calculate sentence probabilities using the language models and compare which model is more suitable. 3) Perform sentiment analysis on music reviews using rule-based and supervised techniques with NLTK and SAS Text Miner, reporting accuracy, precision, recall, and F1 scores. The deliverables are a 2000 word report, Python code, and uploading the work through the LMS. Academic integrity policies are also outlined.

    Original Description:

    txsa group assignment questions

    Original Title

    CT107-3-3-TXSA - Group Assignment (2)

    Copyright

    © © All Rights Reserved

    Available Formats

    DOCX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document outlines a group assignment for a text analytics course. Students will form groups of 3-4 members to complete the following tasks: 1) Build unigram and bigram language models on a provided text corpus to calculate smoothed and unsmoothed probabilities. 2) Calculate sentence probabilities using the language models and compare which model is more suitable. 3) Perform sentiment analysis on music reviews using rule-based and supervised techniques with NLTK and SAS Text Miner, reporting accuracy, precision, recall, and F1 scores. The deliverables are a 2000 word report, Python code, and uploading the work through the LMS. Academic integrity policies are also outlined.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    114 views4 pages

    CT107-3-3-TXSA - Group Assignment

    Uploaded by

    Sharveen Veen
    This document outlines a group assignment for a text analytics course. Students will form groups of 3-4 members to complete the following tasks: 1) Build unigram and bigram language models on a provided text corpus to calculate smoothed and unsmoothed probabilities. 2) Calculate sentence probabilities using the language models and compare which model is more suitable. 3) Perform sentiment analysis on music reviews using rule-based and supervised techniques with NLTK and SAS Text Miner, reporting accuracy, precision, recall, and F1 scores. The deliverables are a 2000 word report, Python code, and uploading the work through the LMS. Academic integrity policies are also outlined.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    of 4

    CT107-3-3-TXSA Group Assignment Page 1 of

    Group Assignment (25% of the total module marks)

    In this part of the assignment, you are required to work on the implementation of text analytics
    techniques and methodologies using Python and SAS Text Miner.

    Group: Form a group comprises of Three or Four members to work on this assessment
    component.

    Use the Text Corpus.txt given in the Group Assignment Data.zip to answer Q1 – Q3 of this
    assessment component.

    The file Text Corpus.txt is a text corpus which contains three sentences. Each sentence is
    limited by the sentence pads such as <s> and </s> as the starting and end of the sentence
    respectively. You are expected to use the appropriate equations to perform the respective tasks.
    The unknown words should be treated as UNK.

    Perform suitable pre-processing to perform the following tasks.

    Q1. Form a unigram language model (25 marks)


    1) Compute manually the unsmoothed unigram probabilities and tabulate the respective
    values for all the given tokens. (6 marks)
    2) Compute manually the smoothed unigram probabilities using the Laplace smoothing
    technique and tabulate the respective values for all the given tokens. (7 marks)
    3) Implement unsmoothed and smoothed unigram language models in python and report the
    output. (12 marks)

    Q2. Form a bigram language model (25 marks)


    1) Compute manually the unsmoothed bigram probabilities and tabulate the respective
    values for all the given tokens. (6 marks)
    2) Compute manually the smoothed bigram probabilities using the Laplace smoothing
    technique and tabulate the respective values for all the given tokens. (7 marks)
    3) Implement unsmoothed and smoothed bigram language models in python and report the
    output. (12 marks)

    Q3. Work on sentence probabilities (25 marks)


    Using the smoothed model values, carry out the following tasks.
    1) Compute manually the sentences probabilities using the unigram model. (5 marks)

    Level 3 Asia Pacific University (APU) 202212


    CT107-3-3-TXSA Group Assignment Page 2 of

    2) Compute manually the sentences probabilities using the bigram model. (5 marks)
    3) Justify which language model is more suitable to calculate the sentence probabilities. (5
    marks)
    4) Implement and report the respective sentence probabilities in python using both unigram
    and bigram language models. (10 marks)

    Q4. Supervised Text Classification (25 marks)


    Use the Musical_Instruments_Reviews.csv data set available in the Group Assignment
    Data.rar to perform the following tasks.

    1. Predict the sentiments (consider 3 levels) for each review found in the data set using the
    Rule-Based Unsupervised Technique (a library named VADER from NLTK). Implement
    using suitable python codes and report the portion of the results. (3 marks)

    2. Export the resulting data set along with the predicted sentiments to a .csv format using
    python and report the relevant code used for this operation. (Note: The resulting data set
    must be submitted via the given submission link in MOODLE). (2 marks)

    3. Use the data set exported in the Q4.2 to build a supervised sentiment classification model
    using Naïve Bayes Classifier from NLTK and report the following model performance
    measures:
    a. Accuracy
    b. Precision
    c. Recall
    d. F1 Score

    The python codes must be neat with clear output. Provide relevant comments in the code
    to explain the purpose of the code snippet.

    Deliverable:

    a. Complete & running Python code.


    b. Output stating the above FOUR (04) performance measures.
    (10 marks)

    Level 3 Asia Pacific University (APU) 202212


    CT107-3-3-TXSA Group Assignment Page 3 of

    4. Use the data set exported in the Q4.2 to build a supervised sentiment classification model
    using SAS Text Miner and report the following model performance measures:
    a. Accuracy
    b. Precision
    c. Recall
    d. F1 Score

    The process flow diagram must be neat and provided in the report. Provide relevant and
    necessary explanations with the suitable output (screenshots) to support the answers.

    Deliverable:

    a. Process flow diagram (.xml)


    b. Output stating the above FOUR (04) performance measures
    (10 marks)

    Deliverables:
    Report:
    Word count: 2000 words

    The report must be prepared in a professional manner following the proper documentation
    formats.

    Codes: Python code should run without any arguments. It should read files in the same directory.
    The output must be as specified in respective questions. Suitable comments must be inserted in
    proper places in the code.

    Softcopy: The relevant softcopies such as the report (.doc or .docx or .pdf) and the python code
    files (.py or. ipynb) must be uploaded via the specified submission links available in the
    MOODLE.

    Academic Integrity
    Copying or paraphrasing someone's work (code included), or permitting your own work to be
    copied or paraphrased, even if only in part, is not allowed, and will result in disciplinary action.
    Your grade should reflect your own work.

    Level 3 Asia Pacific University (APU) 202212


    CT107-3-3-TXSA Group Assignment Page 4 of

    Basically, 'plagiarism' means representing someone else's work as if it is your own. This is a very
    serious academic offence for all students within the University regulations and is particularly
    reprehensible for a researcher. Please do not even consider it. Remember that accidental
    plagiarism (or the appearance of it) may be avoided by referencing your work properly. This
    gains you credit, not loses it! The simple rule is that you must not represent the ideas of other
    people (whether they are published works or the work of other students) as your own.

    Carefully read APU guidelines and policies on the Webspace. The golden rule on plagiarism is
    DO NOT DO IT.

    Level 3 Asia Pacific University (APU) 202212

    You might also like