Correlation and linear regression data analysis

Six Sigma

Excerpt coming from ‘Data Analysis’ chapter:

Correlation, Basic Linear Regression

In this newspaper, we perform a linear regression analysis on previously accumulated data related to the number of daily e-mails received (R) and sent (S) by a particular user (the author). We certainly have depicted the first daily e-mail data as a time series, incrementing D. By 1 for each day’s measurement. The computed regression coefficient l is the slope of the regression line.

A moment series contains sequencing effective data details at consistent time times. As such, this exercise symbolizes a important statistical examination to determine whether a natural provisional, provisory ordering is usually inherent in the data. It ought to be recalled that the collection of data for each of R. And S. contains 15 daily samples, which are collected during two exercises spanning 15 and a few days respectively. This component will be noted in the examination to follow. Table 1 displays the predicted values and subsequent regression analysis intended for the emails received (R). Table a couple of illustrates the predicted beliefs and future regression analysis for the e-mails sent (S).

It summarizes the regression examination for nachrichten received (R).

Time Series (N)

Nachrichten Received (R)

Predicted Value

1

2 . 581

a couple of

0. 147

3

6th. 098

some

2 . 311

5

ninety-seven

6. 909

6

seventy two

10. 290

7

seventy eight

9. 073

8

seventy seven

9. 614

9

87

8. 261

10

93

7. 400.00

11

56

12. 454

12

67

10. 966

13

seventy

10. 561

14

sixty one

11. 778

15

63

11. 507

Table one particular: Predicted Values for R.

Based on the web analysis using the tools simply by Waner et. al. (1999), the geradlinig regression formula, the regression coefficient 3rd there’s r, and the producing graphical characterization of the time series for Ur. are given simply by:

y = f (R) = -0. 135243 back button + 20. 0276

l = -0. 843775

Figure 1: linear regression to get R, most 15 days

Research of the results for variable R. screen a distinct downwards trend over time, and the best-fit regression series does may actually visually correspond to the division of data. This might therefore be described as a credible interpretation of the craze of received e-mails (R), over a period of 12-15 consecutive times.

This section summarizes the regression analysis intended for e-mails sent (S).

Time Series (N)

E-mails Dispatched (S)

Forecasted Value

1

68

six. 771

2

72

almost eight. 280

3

64

7. 263

four

55

6. 118

your five

43

four. 592

6

49

a few. 355

several

52

5. 737

8

55

6th. 118

being unfaithful

46

four. 974

10

62

several. 008

11

98

11. 586

12

12. 730

13

14. 967