Other questions

Question 1

What is sensitivity ?

Answer 1

How far you are right in +VE

Question 2

What is True positive value ?

Answer 2

Sensitivity

Question 3

What is specificity ?

Answer 3

How left you are in -VE

Question 4

What is true negative value ?

Answer 4

Specificity

Question 5

What are the different types of distribution ?

Answer 5

.Normal
.Skewed
.Uniform
.Bi-modal

Question 6

What calculation used to calculate an outlier ?

Answer 6

Q1-1.5xIQR

Question 7

How to handled missing data ?

Answer 7

.Remove the entire sample
.Imputation make a reasonable guess as to what to what it may be

Question 8

What is included in central tendency ?

Answer 8

Mean, median, mode

Question 9

What is variability ?

Answer 9

The spread of data
Variance, standard deviation, range

Question 10

CRISP-DM life cycle

Answer 10

A business centric approach to the data science loop

Question 11

What is host datasheet ?

Answer 11

Allow download as hosted files

Question 12

What is API?

Answer 12

Application program interface
Allows you to retrieve info through predetermined functioning

Question 13

What is web scraping?

Answer 13

Scripts used to pull info from websites

Question 14

What are reals ?

Answer 14

Represented as R, it means 3.3, 3.2 3.1 etc

Question 15

What is structured data ?

Answer 15

Dictionary of keys and values
Allows for easy data retrieved

Question 16

What is quantative data ?

Answer 16

Numerical data, which can either be discrete or continuous

Question 17

What is qualitative data ?

Answer 17

Or categorical data
Doesn’t have an inherit numerical value,

Question 18

Give examples of qualiative data ?

Answer 18

Phone number
Post code

Question 19

What is nominal values ?

Answer 19

Qualiative data
. Hair colour
. Nationality
. Name

Question 20

What are ordinal values ?

Answer 20

Qualitve data
. Has an order
. Likely, not likely, unlikely…

Question 21

What is a simple hypothesis?

Answer 21

State a relation between the dependent and independent variable

Question 22

What is a complex hypothesis ?

Answer 22

State a relationship between multiple dependent and and independent variables

Question 23

What is a directional hypothesis ?

Answer 23

Predicts how independent variables affect the dependent variables

Question 24

What is unstructured ?

Answer 24

Does not conform to a defined model structure

Question 25

What is data ?

Answer 25

A small piece of information A recorded observation

Question 26

What is atomic data ?

Answer 26

Primitive data type Cannot be broken down further

Question 27

What is composite data ?

Answer 27

Can be broken down into atomic or more composite String, lists etc

Question 28

What is normalizing a dataset ?

Answer 28

.Ensures all features contribute equality .Use when the model needs equal weight (k means)

Question 29

What is standardising a dataset ?

Answer 29

.Rescaling data so that it has a mean of 0 and a standard deviation of 1 .

Question 30

What are the two types of variables ?

Answer 30

Categorical and numeric

Question 31

What types of categorical data can you have ?

Answer 31

Nominal, no order (gender) Ordinal, ordered (good - bad)

Question 32

What types of numeric data can you have ?

Answer 32

Continuous, infinite like height Discrete, number of people

Question 33

What is a global optimum ?

Answer 33

Absolute best solution

Question 34

What is the scale of normalization? and when is it best used ?

Answer 34

0-1 When data has a fixed range

Question 35

What is the downside of normalization?

Answer 35

Its sensitive to outliers

Question 36

When is standarization best used ?

Answer 36

Data may go beyond the sample range, handles outliers better

Question 37

Why reduce dimensionality ?

Answer 37

Curse of dimensionality = too many dimensions .Hard to visualize .Slower to computer .Possible irrelevant features

Question 38

What is true a true positive ?

Answer 38

Corrected predicted positive cases

Question 39

What is a ROC curve and what does it plot ?

Answer 39

Helps you see how well a classifier performs across thresholds TPR vs FPR

Question 40

What's the issue with training data to evaluate?

Answer 40

You're testing on data the model has already seen

Question 41

What is underfitting ?

Answer 41

High bias, low variance (too simple)

Question 42

What is overfitting ?

Answer 42

Low bias, high variance (too complex)

Question 43

How to guard against overfitting ?

Answer 43

Split the data: 80% training set 20% test set