[SOLVED] deep learning html python database statistic software network SEMESTER 2 2018/19 COURSEWORK BRIEF:

SEMESTER 2 2018/19 COURSEWORK BRIEF:
Module Code:
MANG 3073
Assessment:
Individual Coursework 2
Weighting:
60%
Module Title:
Analytics in Action
Module Leader:
Cristian Bravo
Submission Due Date: @ 16:00
Method of Submission:
7th June, 2019
Word Count:
2000
Electronic via Blackboard Turnitin ONLY
(Please ensure that your name does not appear on any part of your work)
Any submitted after 16:00 on the deadline date will be subject to the standard University late penalties (see below), unless an extension has been granted, in writing by the Senior Tutor, in advance of the deadline.
University Working Days Late:
Mark:
1
(final agreed mark) * 0.9
2
(final agreed mark) * 0.8
3
(final agreed mark) * 0.7
4
(final agreed mark) * 0.6
5
(final agreed mark) * 0.5
More than 5
0
This assessment relates to the following module learning outcomes:
A. Knowledge and Understanding
A3. Solutions and technologies specifically designed for handling big data.
B. Subject Specific Intellectual and Research Skills
B4. Handle various types of queries with big data sets.
B5. Work with current software packages to create models using complex data sources.
C. Transferable and Generic Skills
C1. Self-manage the development of learning and study skills
C2. Plan and control effectively for successful completion of a personal workload
C3. Communicate effectively.
3
Coursework Brief:
Crowdfunding has been a popular way for many, from people to big companies, to fund risky, uncertain, or otherwise obscure projects. The risk in a crowdfunded project comes from the inability of the project controllers to deliver what they promise; and given the huge rise in low quality projects requesting funding, the ability of the platforms being able to detect when things start to go awry is becoming a necessity.
In this project, you are given a dataset1 of comments arising from two well-known crowdfunding platforms (Kickstarter and IndieGoGo), along with the sentiment of the comment. These comments have been tagged by real people, so there is also a confidence score associated to it. The variables are:
ID: Comment ID (not predictive)
Text: Comment in text format. Scraped from the web.
Sentiment: Either positive (1) or negative (0).
Confidence: How confident was the scorer on the sentiment score.
Starting from this sample, your task is to create a model that infers the sentiment from the text comments and design the way the company will put it into production, using what you know about Deep Learning and Big Data technologies. For this purpose, write a report that answers the following questions:
1. Reflect critically on the required big data capabilities needed for the company to operate. How would you use big data technologies to manage the volume of reviews stored in the site? Comment on what technologies
1 Original data by Doson. https://www.kaggle.com/dosonl/comments-from-funding-platform

SEMESTER 2 2018/19
2. 3.
4.
5.
6.
you believe are necessary (e.g. Hadoop, NoSQL, traditional relational databases, data lakes, data warehouses, etc). Remember to use relevant literature to justify your answer.
Preprocess the text so that it can be used in Deep Learning. Explain your decisions and describe the resulting datasets.
Study the distribution of the dataset and extract basic statistics from the document. At the very minimum, discuss the words that are repeated the most, the length of the phrases, and the distribution of the comments.
Train several Neural Networks to predict whether a comment is speaking in a positive or negative way about a project. Try at least two different embeddings (e.g. fastText, GloVe, BERT, etc.) as input layers, with one of them being training your own embedding; and two different architectures, one of them including more than one hidden layer (i.e. not Kims model). Discuss and justify your choice of the architecture, layers, and other parameters for each of the architectures. Comment on how you think it is best to use the confidence score.
Calculate your accuracy, AUROC, and your confusion matrix over the test set for each of your models. Generate a plot with all the ROC curves of your models. Which embedding works best? Which architectures?
Discuss how you would implement your model into production, referring to technologies, process and people, to improve the success chances of your model.
In terms
discuss your results in a rigorous way! Attach as appendices your Google Colab or Jupyter Notebook printed to PDF.
of software, use Python, and Excel as needed. Carefully report the various steps of your methodology and

SEMESTER 2 2018/19
Nature of Assessment: This is a SUMMATIVE ASSESSMENT. See Weighting section above for the percentage that this assignment counts towards your final module mark.
Word Limit: +/-10% either side of the word count (see above) is deemed to be acceptable. Any text that exceeds an additional 10% will not attract any marks. The relevant word count includes items such as cover page, executive summary, title page, table of contents, tables, figures, in-text citations and section headings, if used. The relevant word count excludes your list of references and any appendices at the end of your coursework submission.
You should always include the word count (from Microsoft Word, not Turnitin), at the end of your coursework submission, before your list of references.
Title/Cover Page: You must include a title/ cover page that includes: your Student ID, Module Code, Assignment Title, Word Count. This assignment will be marked anonymously, please ensure that your name does not appear on any part of your assignment.
References: You should use the Harvard style to reference your assignment. The library provide guidance on how to reference in the Harvard style and this is available from: http://library.soton.ac.uk/sash/referencing
Submission Deadline: Please note that the submission deadline for Southampton Business School is 16.00 for ALL assessments.
Turnitin Submission: The assignment MUST be submitted electronically via Turnitin, which is accessed via the individual module on Blackboard. Further guidance on submitting assignments is available on the Blackboard support pages.
It is important that you allow enough time prior to the submission deadline to ensure your submission is processed on time as all late submissions are subject to a late penalty. We would recommend you allow 30 minutes to upload your work and check the submission has been processed and is correct. Please make sure you submit to the correct assignment link.
You will know that your submission has completed successfully when you see a message stating Congratulations your submission is complete. It is vital that you make a note of your Submission ID (Digital Receipt Number). This is a unique receipt number for your submission, and is proof of successful submission. You may be required to provide this number at a later date. We recommend that you take a screenshot of this page, or note the number down on a piece of paper. You should also receive an email receipt containing this number, and the number can be found after submitting by following this guide. This method of checking your submission is particularly useful in the event that you dont receive an email receipt.
You are allowed to test submit your assignment via Turnitin before the due date. You can use Turnitin to check your assignment for plagiarism before you submit your final version. See Viewing Your Originality Report for guidance. Please see the Module Leader/lecturer on your module if you would like advice on the Turnitin Originality report.
The last submission prior to the deadline will be treated as the final submission and will be the copy that is assessed by the marker.
It is your responsibility to ensure that the version received by the deadline is the final version, resubmissions after the deadline will not be accepted in any circumstances.
Important: If you have any problems during the submission process you should contact ServiceLine immediately by email at [email protected] or by phone on +44 (0)23 8059 5656.
Late Penalties: Further information on penalties for work submitted after the deadline can be found here.
Special Considerations: If you believe that illness or other circumstances have adversely affected your academic performance, information regarding the regulations governing Special Considerations can be accessed via the Calendar: http://www.calendar.soton.ac.uk/sectionIV/special-considerations.html

SEMESTER 2 2018/19
Extension Requests: : Extension requests along with supporting evidence should be submitted to the Student Office as soon as possible before the submission date. Information regarding the regulations governing extension requests can be accessed via the Calendar: http://www.calendar.soton.ac.uk/sectionIV/special-considerations.html
Academic Integrity Policy: Please note that you can access Academic Integrity Guidance for Students via the Quality Handbook: http://www.southampton.ac.uk/quality/assessment/academic_integrity.page?. Please note any suspected cases of Academic Integrity will be notified to the Academic Integrity Officer for investigation.
Feedback: Southampton Business School is committed to providing feedback within 4 weeks (University working days). Once the marks are released and you have received your feedback, you can meet with your Module Leader / Module Lecturer / Personal Academic Tutor to discuss the feedback within 4 weeks from the release of marks date. Any additional arrangements for feedback are listed in the Module Profile.
Student Support: Study skills and language support for Southampton Business School students is available at: http://www.sbsaob.soton.ac.uk/study-skills-and-language-support/.
External Examiner:
External Examiner Comments:
Final Approval by External Examiner Date:
Module Leader Response to External Examiner:
(Please note these comments are REQUIRED and will be sent to the External Examiner)