Data And Analysis

Front 1

Key

Back 1

A set of attributes is a key if it is a minimal set of identifying attributes - removing any one attribute would make it no longer uniquely identifiable.

Front 2

Super Key

Back 2

A set of attributes is a superkey for an entity if those attributes, taken together, always uniquely identify every entity instance.

Front 3

Composite Key

Back 3

A Primary Key made up of multiple attributes.

Front 4

Primary Key

Back 4

A single key that is unique and not-null. It is one of the candidate keys.

Front 5

Candidate Key

Back 5

A candidate key can be uniquely and used to identify a database record. They are Not Null and Unique.

Front 6

Representative Sample

Back 6

The corpora should contain a similar mix of text to the language variant for which it is being developed. For example using Shakespeare's works is not a good representation of Elizabethan English.

Front 7

Finiteness

Back 7

A copora should be finite. When building a corpora it is usually decided at the outset how the language is to be sampled and how much data to include.

Front 8

Monitor Corpora

Back 8

These capture the growth and change of a language. They remain finite but extend over time.

Front 9

Name 4 distinctive features of a Machine Readable Corpus compared to books with printed text

Back 9

• They can be huge in size; up to a billion words

• They can be searched and analysed efficiently

• They can be made available to many users simultaneously, at large distances

• They can easily (and sometimes automatically) be annotated with additional useful information.

Front 10

What is the advantage of using a standard reference?

Back 10

Having a standard reference allows competing theories about the language variety to be compared against each other on the same sample data.

Front 11

Name 4 notable English Language Corpora

Back 11

• Oxford English Corpus (OEC)

• Corpus of Contemporary American English (COCA)

• British National Corpus (BNC)

• Brown Corpus

Front 12

Balancing

Back 12

Balancing ensures that the linguistic content of a corpus represents the full variety of the language sources for which the corpus is intended to provide a reference. For example a balanced text corpus includes materials from sources such as books, newspapers, magazines, letters, etc.

Front 13

Sampling

Back 13

Sampling ensures that the material is representative of the types of source. For example, Sampling from the newspaper text involves selecting texts randomly from different newspapers, issues and sections.

Front 14

List some of the "dimensions" of the source material that balancing would affect.

Back 14

• Language Type: Editied Text, Spontainious, Scripted

• Genre

• Domain (What is the text about?)

• Medium

Front 15

Tokenization

Back 15

Tokenization divides are textual data into tokens such as words, numbers and punctuation marks.

Front 16

Sentence Boundary Detection

Back 16

Identify the start and end of individual sentences.

Front 17

Why Annotate A Corpus?

Back 17

It adds informstion to the corpus that is not explicit in the data itself. This is often specific to a particular application; and a single corpus may be annotated in multiple ways.

Front 18

Annotation Scheme

Back 18

Annotation Scheme is the basis for annotation, made up of a tag set and annotation guidelines

Front 19

Annotation Guidelines

Back 19

Tells annotators - domain experts - how a tag set should be applied. It ensures consistency across annotators.

Front 20

Tag Set

Back 20

An inventory of labels for makeup.

Front 21

Relationship

Back 21

A relationship is an association between entities.

Front 22

Relationship Instance

Back 22

Each individual occurrence of the relationship is a relationship instance.

Front 23

Relationship Set

Back 23

A collection for all instances of a relationship

Front 24

What is this in an ER diagram?

Back 24

Attribute

Front 25

What sort of key is used to identify Course?

Back 25

Composite Key

Front 26

What is Atomicty?

Back 26

All or Nothing: a transaction eitehr runs to completion, or fails and leaves the database unchanged.

This may involve a rollback mechanism to undo a partially-complete transaction

Front 27

What is Consistency?

Back 27

Applying a transaction in a valid state f the database will always give a valid result state.

Front 28

What is Isolation

Back 28

Concurrent transactions hav eht esame effect as sequential ones: the outcome is as if they were done in order.

(NOTE: Transactions may, in fact run at the same time: but should never see each other;s imtermediate state.)

Front 29

What are the ACID properties amd why are they meeded/

Back 29

The ACID properties (Atomicity, Consistency, Isolation, Durable) are a key benchmark for assessing database systems.

Front 30

What is the symbol for total participation in an ER diagram?

Back 30

Double lines

Front 31

Which way should an arrow face in a one to many relationship in an ER diagram?

Back 31

The arrow should point from the many to the one. It should go from the many to the relationship block.

Front 32

What is a weak entitiy?

Back 32

A weak entity is an entity that has attributes but may not be enough to uniquely identify its self without it's identifying relationship and identifying owner.

Front 33

In a tree of an XPath data model, what do the positions of the nodes relative to each other show?

Back 33

They show where they appear in the XML document. Those appearing first in the document will appear leftmost in the diagram.

Front 34

In a DTD declarations, what do the order of the lines mean?

Back 34

Nothing! As DTD is declarative the lines can appear in any order.

Front 35

What would the ELEMENT line for this node look like? *Insert publisher node from paper*

Back 35

<!ELEMENT publisher (name,imprint+)>

The publisher may have only one name, however the plus indicates that there maybe more than one imprint.

Front 36

In a DTD document, when would you use the tag #PCDATA?

Back 36

PCDATA is used for text nodes in DTD.

Front 37

What would the attribute line(s) look like for this node?

Back 37

<!ATTLIST book code CDATA #REQUIRED>

<!ATTLIST book type (hardback|paperback) "paperback">

Front 38

When is CDATA used in a DTD document?

Back 38

CDATA is used to represent an open attribute (no regex)

Front 39

When is #REQUIRED used in DTD

Back 39

#REQUIRED is used when the attribute of a node cannot be left empty.

Front 40

When should not null be used in a SQL schema?

Back 40

Not null should be used on non-primary key attribute if it cannot be left empty or on a foreign key as this must exist for a relationship to exist.

Front 41

Do primary keys require the not null tag in a SQL schema?

Back 41

No. As primary keys by nature are required to make a record in a database thus are not null by default.

Front 42

How do you create a composite key in a SQL schema?

Back 42

primary key (field1, field2,...)

Front 43

How do you create a foreign key in a SQL schema?

Back 43

foreign key x references table(y)

Front 44

How is text represented in a SQL schema?

Back 44

VARCHAR(x) where x is the number of characters

Front 45

In an SQL query, what does distinct mean?

Back 45

distinct is used after select to remove duplicate entries in the result.

Front 46

How do you create a SQL query that involves more than one table?

Back 46

select table1.field

from table1, table2

where table1.field = table2.field and ...

Note: field may be called different things in different tables

Front 47

What is catagorical data?

Back 47

Categorical data/scale is data that has no numerical or natural order.

Front 48

What is ordinal data?

Back 48

Ordinal data/scale give a recognised order between data items, but there is no arithmetic content. Numbers may still be used but there is no way to apply arithmetic to them.

Front 49

What is an interval scale?

Back 49

An interval scale assigns a numeric balue to data, but where these values are relative to each other. Values can be compared, averaged and subtracted but not added together or multiplied.

Front 50

What is a ratio scale

Back 50

A ration scale uses numeric values which have an abstemious notion of zero This means they can sensibly be added, and multiples by real numbers.

Front 51

Give 3 examples of categorical data

Back 51

Classifying words (nouns, verbs, adjective)

Eye Color

Places in a town

Front 52

Give 3 examples of ordinal data

Back 52

T-Shirt Size (XS,S,M,L)

1st, 2nd, 3rd

Shoe size

Front 53

Give 2 examples of interval data

Back 53

Times Of Day

Temperature (Celsius)

Front 54

Give 3 examples of ratio data

Back 54

Temperature (Kelvin)
Wind Speed

Height

Front 55

What is the formula for the mean?

Back 55

*insert picture*

Front 56

What is the formula for the standard deviation with large data sets?

Back 56

*INSERT PICTURE(

Front 57

What is the formula for the standard deviation for a population

Back 57

*insert picture* This is known as the Bessel Correction

Front 58

How do you calculate the median?

Back 58

Order the data (if possible) from smallest to large then find the value that lies in the (n+0.5)/2 position. If the n+0.5/2 value is not an integer then find the value between the (n/2) and (n+1)/2 value.

Front 59

How do you calculate the mode?

Back 59

Count the most frequent value.

Front 60

What is the cosine formula

Back 60

insert picture

Front 61

What is Precision?

Back 61

Precision is what proportion of the documents returned by the system are relevant

Front 62

What is Recall?

Back 62

What proportion of all the relevant documents are returned by the system.

Front 63

Name each section of the contingency table:

*picture*

Back 63

a = true positives

b = true negatives

c = false positives

d = true negatives

Front 64

How do you calculate precision?

Back 64

TP/ (TP+FP)

Front 65

How do you calculate recall?

Back 65

TP/(TP+FN)

Front 66

What is the F Score of a system?

Back 66

F Score is a measure of how balanced a system is, a score close to one has more weight on precision, and the closer to zero it gets the more weight on recall.

Front 67

How do yo calculate F Score?

Back 67

(2PR)/(P+R)