The objective of this assignment is to scrape consumer reviews from a set of web pages and evaluate the performance of text classification on the data. The reviews have been divided into five categories here:
Each review has a star rating. For this assignment, we will assume that 1-star to 3-star reviews are “negative”, and 4-star to 5-star reviews as “positive”.
The assignment should be implemented as a single Jupyter Notebook (not a script). Your notebook should be clearly documented, using comments and Markdown cells to explain the code and results. The assignment can be completed either individually or in pairs.
In this assignment you should complete all of the following tasks:
Select two review categories of your choice. Scrape all reviews for each category and store them as two separate datasets. For each review, you should store the review text and a class label (i.e. whether the review is “positive” or “negative”).
For both category datasets:
From the reviews in this category, apply appropriate preprocessing steps to create a numeric representation of the data, suitable for classification.
Build a classification model using a classifier of your choice, to distinguish between “positive” and “negative” reviews.
Test the predictions of the classification model using an appropriate evaluation strategy. Report and discuss the evaluation results in your notebook.
Evaluate how well your two classification models transfer between category. That is, run experiments to:
Train a classification model on the data from “Category A”, and evaluate its performance on the data from “Category B”.
Train a classification model on the data from “Category B”, and evaluate its performance on the data from “Category A”.
The assignment can be completed either individually or in pairs. Any evidence of plagiarism will result in a 0 grade.
For the assignment, only these third-party packages can be used: NumPy, Pandas, Scikit-learn, NLTK, SciPy, Requests, BeautifulSoup, Matplotlib, Seaborn, Gensim.
The ready solutions purchased from Library are already used solutions. Please do not submit them directly as it may lead to plagiarism. Once paid, the solution file download link will be sent to your provided email. Please either use them for learning purpose or re-write them in your own language. In case if you haven't get the email, do let us know via chat support.
Get Free Quote!
259 Experts Online