**MN3515 Moodle Quiz Assessment**

**What do I do if I have a problem during the quiz?**

If your Internet connection suddenly disappears (as with Virgin this week), or Moodle stops working, what do you do? You let us know as soon as you can, telling us exactly what happened, when and for how long (be as specific as possible and include screen shots if you can of any error messages or other proof of what you are saying) and then we will decide what should happen next. Email to: C.Harbor@rhul.ac.uk (as College has awarded an extra day’s holiday to most staff on 11^{th} May, I don’t know who else might be working).

You will need a laptop or PC to do this quiz – you won’t have time to do it on a phone, iPad/tablet or anything without a proper keyboard. If this is going to a problem, let us know now.

**How much time do I have?**

You have 90 minutes to complete the quiz and submit it. If you have registered and have a statement of special educational needs, then you will get additional time. If the time runs out before you have submitted, then I have set the quiz so that it will be automatically submitted for you.

Although this exam is inevitably open book, you won’t have time to look up all the answers in your notes or the textbooks. Also, many of the questions are too specific for you to be able to look them up – they are not just about regurgitating knowledge but about applying it. So please do revise using materials available on the Moodle page for MN3515:

- Lecture slides
- Lecture videos (from the RePlay section on the right side of the Moodle page, click on [Show All] to see the complete list
- Workshop notes – especially for examples of the types of tasks you can use for each modelling technique
- Required Reading from the textbooks, especially for the different modelling techniques, and for data understanding and data preparation tasks
- Testing your Understanding Solutions PDF file (last topic on Moodle page) – especially for how to do calculations

**How many questions?**

There are 10 text-based questions which are exactly like those you would get in Part A of a normal exam; each is worth 20 points. For the non-calculation ones, you should expect to write a short paragraph for each – 2 or 3 sentences. The calculation questions will probably take you a bit longer than the others.

There are 5 multiple choice and/or true/false questions; each is worth 5 points

To make it less easy for you to share your answers (cheat), I have put in loads of extra work and made a question bank from which your questions will be selected – so not everyone will get exactly the same questions. However, everyone will get the same number of the same types of question. So, for instance, everyone will get the same number of calculation questions which will be of the same types, but the numbers/data will be different. It is not the numbers which make these calculations difficult but knowing what to do with them, so do not think that your questions are more or less difficult than anyone else’s are.

There will **not** be any questions which require you to write any R code. There will **not** be any questions on Tableau or Excel.

If you got everything right, you would get 100%, but I am not really expecting that, so don’t worry too much if you don’t have time to answer all the questions. If you get stuck on one question don’t waste loads of time trying to figure it out, go on to the next and come back if you have time.

**Who is doing the marking?**

For the 10 text-based questions, I am doing the marking and these questions will then moderated by one of my academic colleagues.

For the 5 multiple choice true/false questions, Moodle is doing the marking and they will not be moderated.

**How do I show working for calculations?**

For calculations you **must** show your working as I am awarding marks for the method as well as a correct answer. This is not too difficult. You will only need a few mathematical symbols but you should locate them in advance on the keyboard you will be using:

+ - = which are just as you expect

* for multiplication

/ for division

() to show how to group parts of your calculation, for instance (3+4)/2 shows that you add 3 to 4 to get 7, and then divide 7 by 2.

You will also need to be able to make a superscript 2 to show squared numbers for calculations such as SSE and SST. To do this within the Moodle quiz question:

- Click
**Show/hide advanced buttons**on the button bar at the top of the text window - When you need to add a superscript 2 to show squared, either
- Type the formula up to where you need the superscript 2 to go, click
**Superscript**, type 2, then press the space bar to get back to normal text

- Type the formula up to where you need the superscript 2 to go, click
- Type the entire formula, go back and click and drag over any 2s which need to be in superscript and then click
**Superscript**(this is a bit more risky, as you are likely to miss some out)

- Type the entire formula, go back and click and drag over any 2s which need to be in superscript and then click

You can practice typing calculations in the Practice Test I have made (it has one non-question) which is on the Moodle page for MN3515 just below where you submitted your coursework assignment. You can access this as many times as you want – but once will probably be enough.

You can use any calculator you want to work out the answers. You need not use more than three decimal places.

**What should I revise?**

If you look at past papers (see instructions I sent you earlier on how to access past papers when you are off campus), then you should have a very good idea of what to revise (I will not give you solutions to past exam papers so don’t ask for them – but if you look at the video of the exam revision workshop in the last topic of the Moodle page for MN3515, we discussed answers to some of the questions from last year’s paper). You will find information in lectures, workshop notes, and the readings from the textbooks (which you should have done already). Below are some notes:

- You need to know about
**all**the modelling techniques we used during the course in R- Their main features, concentrating on what is unique about them
- Disadvantages and advantages
- Examples of the types of tasks you could use them for

- The types of questions we looked at, including calculations, in the Testing your Understanding slides in lectures. You can find solutions for all of these in a PDF file which is in the last topic on the Moodle page for MN3515
- The types of problems you might encounter with data, why they are a problem and how you might deal with them (not the R code, but the general technique). You should have done some of this in the data preparation phase of your assignment.
- Outliers
- Null values – Missing values
- Duplicate records
- Etc.

- The purposes and types of information you might discover in the data understanding phase, what they might tell you, how you might deal with revealed problems
- Correlation
- Multicollinearity
- Etc.

- You need to know the meaning of various technical terms, how they are used, why, what they tell you, etc. Some terms have synonyms, and you need to know them all. You need to know how to calculate the ones which you can calculate. The following is not necessarily an exhaustive list:
- Supervised data mining models
- Unsupervised data mining models
- Outcome variable – Dependent variable – Response variable – Target variable (synonyms)
- Variable – Feature – Attribute (synonyms)
- Feature vector (a set of variables/features useful for your analysis)
- Use phase (how would you implement your model in practice)
- Instance – Observation (synonyms)
- Linear regression
- Baseline prediction
- Sum of Squared Errors (SSE)
- Total Sum of Squares (SST)
- R
^{2}

- Logistic regression
- False positive
- False negative
- MAE
- RMSE
- ROC curve
- Threshold value
- AUC
- Confusion matrix
- Specificity – True negative rate (synonyms)
- Sensitivity – Recall – True positive rate (synonyms)
- Accuracy

- Association rule mining
- Support
- Confidence
- Lift

- Clustering
- Collaborative filtering
- Content filtering

- Generalisation
- Overfitting
- Underfitting