Week 2

Question 1

5 steps in Machine Learning Process

Answer 1

• Data Collection
• Data Exploration and Preparation
• Model Training
• Model Evaluation
• Model Improvement

Question 2

Data Collection

Answer 2

Involves gathering learning materials or data set that algorithm will use to generate an actionable knowledge or intelligence. In most cases, the data will need to be combined into a single source like a spreadsheet or data set.

Question 3

Data Exploration and Preparation

Answer 3

Exploring the data generally involves getting a basic understanding of a dataset through numerous variable summaries and visual plots. Through this exploration, we will often identify problems with the data, including missing values, noise, erroneous data, and skewed distributions.

Question 4

Model Training

Answer 4

A subset of the entire data set and is used to build up the model.

Question 5

Model Evaluation

Answer 5

• Because each machine learning results in a biased solution to the learning problem it is important to evaluate how well the algorithm learns from its experience and by the experience of the algorithm as more data becomes available.
• Its the stage in the machine learning process where you test how well the algorithm has work from the data

Question 6

Model Improvement

Answer 6

If better performance is needing you might need to supplement data to become more efficient or perform additional steps to acquire accurate data

Question 7

Vectors

Answer 7

• Fundamentals data structure of R
• all the elements in the vector must be of the same type
• either all elements are ‘characters’ or all elements are numeric or logical
• two special values “Null” and “NA” to indicate missing values
• R vectors are ordered so accessing it would require counting the position of the element
  
  • Indexing always begins with [1]

Question 8

How are Vectors created?

Answer 8

Vectors are created using

c()

Example :

subjects

Question 9

Factors

Answer 9

• A special case of the vector that is solely used to represent categorical or ordinal variables.
• Example: Creating a factor from a vector
  • subjects
  • Faculty

Question 10

Categorical vs Ordinal Variables

Answer 10

1. The categorical or nominal variable has one or more categories. (female, male)
2. The ordinal variable has a clear ordering of its elements
   
   • (high, medium, low)

Question 11

List

Answer 11

• Used for storing an ordered set of elements
• unlike vecotor, list is a collection of all kinds of data elements
  
  • subject1
  • temperature = temperature[1],
  • flu_status = flu_status[],
  • gender = gender[1],
  • blood = blood[1]

Question 12

Data Frame

Answer 12

• Most used data structure in R
• List of vectors or factors, each having exactly the same number of values
• syntax: data.frame()
• example:
  
  • emp.data
  • emp_id = c (1:3),
  • emp_name = c(“Pat”, “John”, “Mike”)
  • stringAsFactors = FALSE
  • )

Question 13

Data Frame Contd….

Answer 13

• Analogous to spreadsheet
• a data frame is two dimensional and is often displayed as a matric (rows and columns)
• in the language of machine learning, columns are often called as ‘Features’ or ‘Attributes’, while rows are called ‘examples’

Question 14

to extract a column :

Answer 14

Emp.data$emp_name

Emp.data[c(“emp_name”,”salary”)]

Question 15

Matrix

Answer 15

• Two dimensional table with rows and columns
• matric can contain only one type of data elements
• the syntax for creating a matrix: matric()
• you also need to specify the number of rows and columns

Question 16

Creating a Matrix example

Answer 16

m

this creates a 2x2 matrix

[,1] [2]

[,1] 1 3

[,2] 2 4