Lecture 08: Quality and Sentiment Analysis Flashcards

1
Q

What is quality?

A

Quality is through the use of products and services the ability to satisfy customers and the intended and unintended relevant interested parties.
Quality is:
- Not absolute: different meanings / different situations
- Multidimensional: many contributing factors, not easy to summarise or quantify
- Subject to constraints
- Acceptable compromises

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What are the five approaches to view quality?

A
  1. Transcendent
  2. Product-based
  3. User-based
  4. Manufacturing
  5. Valued-based
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What is the Transcendent approach to view quality?

A

The classical definition that quality is an innate excellence, through uncompromising standards and high achievement, it is impossible to quantify and is difficult to apply in a meaningful sense.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What is the Product-based approach to view quality?

A

The view that quality is precise and measurable - differences can be reflected in the variation of quantity of an attribute. The higher the quality, the higher the cost.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What is the User-based approach to view quality?

A

The view that quality is subjective to the eye of the beholder, people have differing wants and needs, as a result it is highly subjective and not definable or measurable. This creates two problems:

  • How to aggregate varying individual preferences?
  • How to distinguish attributes that imply quality from those that maximise satisfaction?
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What is the Manufacturing-based approach to view quality?

A

The view that quality can be measured based on how the product or service produced conforms to the set of initial requirements for them. Any deviation from these requirements implies a reduction in quality, has an emphasis on:

  • reliability engineering
  • statistical control
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What is the Value-based approach to view quality?

A

The view that quality is as:
- Performance at an acceptable price.
- Conformance at an acceptable cost.
It is about blending two distinct concepts:
- Quality: the measure of excellence
- Value: the measure of worth
The ability provide what a customer requires at a price they can afford.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What is Quality Management as a process?

A

Inspection –> Quality Control - Quality Assurance –> Total Quality Management (TQM)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

In Quality Management, what does the process of Inspection involve?

A

The process of Inspection involves deploying people to inspect in order to ensure a basic level of quality. This process will ensure that standards are met and you can identify areas which do not conform to standards and take corrective action.
It is a reactive process of detection, rather than preventative, and is rather expensive, inefficient and ineffective in comparison to other processes in quality management.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

In Quality Management, what does the process of Quality Control involve?

A

The process of Quality Control is about quality panning and procedures as well basic statistics and performance data. It involves using statistical techniques such as sampling to make decisions. This again a reactive process, and is used after the process of inspection, and can remove chronic problems

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

In Quality Management, what does the process of Quality Assurance involve?

A

The process of Quality Assurance involves developing quality systems and quality planning to ensure a process or product fulfills quality requirements. It is aimed at developing procedures which avoid mistakes, and focuses on prevention rather than detection.
Places an emphasis on:
- Procedure compliance
- Product conformity
These being achieved through product and operations management tracking

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

In Quality Management, what does Total Quality Management (TQM) involve?

A

The process of a centred and management approach to manage quality. The idea that it is the responsibility of all individuals within an organisation, not just managers, to ensure the long term success and benefits of customer satisfaction.
It relies heavily on employee involvement, collaboration and teamwork to achieve that overall goal.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What is the ISO 25012 / 2008?

A

Defines a general data quality model for data retained in a structured format within a computer system. It can be used to:

  • establish data quality requirements
  • define data quality measures
  • Plan and perform data quality evaluations
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

What are the 15 classes of quality attributes?

A
  1. Accuracy
  2. Completeness
  3. Consistency.
  4. Credibility.
  5. Currentness
  6. Accessibility
  7. Compliance
  8. Confidentiality
  9. Efficiency
  10. Precision
  11. Traceability
  12. Understandability
  13. Availability
  14. Portability
  15. Recoverability
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

What is Information Quality?

A

Information which consistently meets all knowledge worker and end-customer expectations.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What is Information Quality Management?

A

Information Quality Management is the application of sound quality principles and processes to information as a product of business and manufacturing processes.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

What is the Information Quality Management process?

A
  1. Assess product and process quality
  2. Control processes that produce the product
  3. Improve processes that product the product to meet or exceed customers’ expectations and requirements.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

What are three categories of information quality?

A

1, Definition and information architecture - data and information are “products”, so they require “product specifications”.

  1. Content - the actual values in a database must represent facts about the real world: completeness and accuracy.
  2. Presentation - user needs timely access to relevant information so they can perform work effectively and efficiently.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

What is data conformance?

A

Data values are consistent with the attribute definition

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

What is data completeness?

A

Each process or decision has all the information it requires. Data can be complete in two-ways:

  1. Record Completeness: a record exists for every real world object or event that should have a value.
  2. Value Completeness: data element has a value stored for all records that should have a value.
21
Q

What is data validity?

A

Values conform to the information product specifications. There are three types of data validity:

  1. Value validity - a data value is valid for this data element
  2. Business rule validity - data values conform to specified business rules.
  3. Derivation validity - derived data value is produced correctly according to a calculation formula or set of derivation rules.
22
Q

What is data accuracy?

A

Data values are correct.
There are two types of data accuracy:
1. Accuracy to surrogate resource - data agrees with original corroborative source.
2. Accuracy to reality - data correctly reflect characteristics of real-word object.

23
Q

What is data precision?

A

Data values are correct to the right level of detail.

24
Q

What is data non-duplication?

A

There is only one record in a database representing a given real-world object or event.

25
Q

What is data equivalence?

A

Data in one database is semantically equivalent to data about the same object or event in another database.

26
Q

What is data concurrency?

A

Concurrent data queries to each database produce the same result.

27
Q

How can you measure information quality?

A

The quality of information can be measured electronically through software.
The quality of information can be measured through physical comparison of data with the real-world object.

28
Q

What are the 9 IAIDQ: Areas of Best Practices?

A
  1. Data Stewardship
  2. Data and Quality Standards
  3. Organisational Issues
  4. Operations and Processes
  5. Data element development and specification
  6. Data management and data quality tools
  7. Measurement
  8. Individual support
  9. Privacy
29
Q

What is the Area of Best Practice - Data Stewardship about?

A

Data Stewardship is about maintaining a corporate program which understands the roles and responsibilities in data:
- Ownership
- Acquisition
- Quality Assurance
- Storage
- Distribution
Make each functional area responsible for their own performance and data management

30
Q

What is the Area of Best Practice - Data and Quality Standards about?

A

Data and Quality standards is about developing internal standards and seeking useful external standards, harmonising them together and promoting consistent operations from a range of systems and processes.

31
Q

What is the Area of Best Practice - Organisational Issues about?

A

Organisational issues is about developing an in-house unit to create and assess data acquisition, and tap into external resources where appropriate.

32
Q

What is the Area of Best Practice - Operations and Processes about?

A

Operations and processes is about developing operations and processes which maximise data quality through the use of state of the art technology to manage data. It also involves monitoring regular requirements which may affect data quality.

33
Q

What is the Area of Best Practice - Data Element and Specification about?

A

Data element and specification is about designing and maintaining data, systems and reporting mechanisms to provide sound data management and quality end-user service.
Review data to inform system development, and support convenient modification and updates.

34
Q

What is the Area of Best Practice - Data Management and Data Quality Tools about?

A

Data management and data quality tools is about developing tools such as corporate data dictionary, business rules, data flow documentation and leverage technology resources for predictive and data visualisation tools

35
Q

What is the Area of Best Practice - Measurements about?

A

Measurements is about developing performance metrics to measure the effects of poor data quality on: correcting errors, writing reports, investigating and preventing errors and regulatory scrutiny. Creating benchmark results for each data resource.

36
Q

What is the Area of Best Practice - Individual Support about?

A

Individual support is about providing institutional support for data management and data quality on an individual level and organisational level.

37
Q

What is Area of Best Practice - Privacy about?

A

Privacy is about educating users about privacy issues, policies and compliance with privacy regulations, as well as areas around controlling access and promoting best practices.

38
Q

What is Information?

A

Information is an organised collection of processed data which has meaning and context, which then goes onto focus on semantics and structured data.
Information is evaluated and a decision is made on that basis of the evaluation of the information, and the information itself, it is used to support a decision with an optimal outcome.

39
Q

What is objective information?

A
  • Unbiased
  • Not touched
  • Verifiable by facts or mathematical calculations
    Facts are a form of objective information, if facts do not exist, then a set of balanced opinions must be used.
40
Q

What is subject information?

A
  • Biased
  • Touched
  • Not verifiable
    Opinions are a form of subjective information, they are always biased and if they are not balanced it is propaganda.
41
Q

What is classification?

A

Classification is the mapping of an object to one or more classes where a class can be described by some properties or attributes.

42
Q

What is machine learning?

A

Machine learning is how to automatically learn to make accurate predictions on past observations.

43
Q

What is the Sentiment Analysis?

A

Classification task where each category represents a sentiment. Sentiments found within comments and feedback provide useful indicators for many different purposes. Need to find relevant sources, extract related sentences with opinions, read them, summaries them, organise them into usable forms

44
Q

How can sentiments be classified?

A

Two categories – positive, negative
N point scale – v good, good, satisfactory, bad, v bad
Full range of emotions

45
Q

Problems with sentiment analysis?

A

Resource consuming task, huge volume of opinionated text, opinions are often buried in long forum posts of blogs

46
Q

What is opinion mining?

A

Grows out of need for automated opinion discovery and summarising systems. The computational study of opinions, sentiments and emotions expressed in text.

47
Q

What is document level sentiment classification?

A

Classifying an opinionated document as expressing a positive or negative opinion

48
Q

What is sentence level sentiment classification?

A

classifying a sentence as subjective or objective.

49
Q

What two tasks does sentence level classification require?

A

Subjectivity classification – determine whether it is a subjective sentence or an objective sentence. Then Sentiment classification – if sentence is subjective, determine whether it expresses a positive or negative opinion