regression analysis example

REGRESSION ANALYSIS

USING BOWLING PIN FACTORY(FACTISM)

IME 435/L DESIGN OF EXPERIMENT

YOUNG KIM

2/25/2014

IME DEPARTMENT

CALIFORNIA STATE POLYTECHNIC UNIVERSITY

Title:
Simple regression analysis using bowling pin factory (FACTISM)

Statement of the problem:
How change in pressure in 8 different levels will affect response, weight, will be investigated by performing simple regression analysis. Some of questions to be answered at the end of the experiment are:

a.      If the chosen factor is correlated with the response variable (launch distance) or not.

b.      Is the correlation positive or negative?

c.       Is the correlation strong or weak?

d.      What is the correlation coefficient? What is its significance?

e.      What is the coefficient of determination? What is its significance?

Materials or Tools:
FACTISM
Microsoft Excel
Microsoft Words

Assumptions:
1. Coefficient of correlation with value of less than 0.85 is considered to be insignificant.
2. Coefficient of determination with value of less than 0.723 is considered to be insignificant.

Procedure:
1. Collect responses from Factism by changing pressure value to 80, 81, 82, 83, 84, 85, and 86 with 6 replications for each level.
2. Perform simple linear regression analysis manually and using excel.

Results:
Using manual method:

 

Findings & conclusion:

a.      If the chosen factor is correlated with the response variable (launch distance) or not.
- From the regression analysis we can see that there is no correlation between the pressure and weight. This is supported by the fact that our slope or b_1 is relatively small. Yet, slope by itself does not tell us the strength of the correlation.

b.      Is the correlation positive or negative?
-Since our slope is positive, the relationship is positive relationship.

c.       Is the correlation strong or weak?
-This correlation is said to be very weak due to its small value in coefficient of correlation.

d.      What is the correlation coefficient? What is its significance?
-The correlation coefficient in this particular analysis is 0.069144. From our assumption, we have decided that R value with less than 0.85 is to be insignificant. Hence, it is confident to conclude that pressure does not have any effect on weight.

e.      What is the coefficient of determination? What is its significance?
-The coefficient of determination in this particular analysis is calculated to be 0.004781 by which mean, any variation in weight is caused by pressure only .5% of the time.

Two factor ANOVA example

TWO FACTOR ANOVA

USING BOWLING PIN FACTORY(FACTISM)

IME 435/L DESIGN OF EXPERIMENT

YOUNG KIM

2/19/2014

IME DEPARTMENT

CALIFORNIA STATE POLYTECHNIC UNIVERSITY

Title:
Two Factors ANOVA using bowling pin factory (FACTISM)

Statement of the problem:
Using statistical analysis of variance, ANOVA, using two factors out of 5, temperature and alloy, experiment will be conducted to see if any of those factor contributes to change in response.

Materials or Tools:
FACTISM
Microsoft Excel
Microsoft Words

Assumptions:
1. Other factors other than temperature and alloy will not have effect on response.
2. H_0 =  u1 =  u2  =  u3  =  u4  =  u5, temperature has no effect on weight
    H_a , temperature does have effect on weight

    H_0 =  u1 =  u2  =  u3  =  u4  =  u5, alloy has no effect on weight
    H_a , alloy does have effect on weight

    H_0 =  u1 =  u2  =  u3  =  u4  =  u5, interaction has no effect on weight
    H_a , interaction does have effect on weight
3. Alpha = 0.95

Procedure:
1. Collect responses from Factism by changing temperature and alloy value to 0, 25, 50, 75, 100, and 30, 35, 40, 45, 50 respectively.  Repeat each trial 8 times.
2. Perform two factor ANOVA by both manual and using Megastats.

Results:
Using manual method:

Performing two way ANOVA manually :

Using Megastats:

 

Findings & conclusion:

Since our calculated F value for temperature is less than our critical F value, we fail to reject our null hypothesis. Our p-value from Megastat also indicates that we fail to reject our null hypothesis since 0.2743 is greater than 0.05 which is our alpha value. However, we reject null hypothesis for alloy since its calculated F value is greater than critical F value, yet, without confident since both p values are so close to each other. Finally we can confidently say that we fail to reject our null hypothesis for interaction of temperature and alloy due to greater magnitude in calculated p value from critical p value.

DOE+LP

So, for my senior project, we are working on process improvement and standardization on current processes. First, we are investigating what is causing the variation in the weighing process. To do that, I will use design of experiment with 2 factors with 3 levels. Hopefully, with some constraints, I would be able to come up with an objective function, which would be prediction equation for DOE, and apply LP method to maximize or minimize that objective. I will post the results asap I have it. 

random thought

Okay.. water level is rising in sea.... we are not getting enough rainfall... oil is getting scarce..

why don't freaking million's engineer work on a project to convert sea water into fresh water at the same time generate electricity by somehow converting energy created by waves or tides..... so we can sustain our sea water level + save gas + get more source of fresh water.... just charge agricultural people more money for use of those water to fund those project and operation... may sound crazy.. but could be our only option in future :(... I would definitely loved to be involved in such a life changing project like this... :) 

sustainability pt1

While I was taking class yesterday, for no reason... this whole scenario came across my mind..

There are two jobs.. each job requires one worker..

Job 1 requires 8 hours... but can be crammed to 7 hours if additional $10 is paid..

Job 2 requires 1 hours...

and hourly wage is $10....

What would you do if you were the boss...

have one worker do both job by cramming Job 1

or 

Have one part time worker to do Job 2.... 

The cost is going to be same in both scenario... which is .. $90... I am sure that there are people who would choose second option because they just can't do simply math and thinks first option would cost them more...

although if you consider all the indirect costs associated with each employees, you will find out that you are actually saving money if you go with first option....

btw...

what you, as an employer, should consider more is the social aspect of sustainability.. which in this case is the motivation of the worker or morale. 

to be continued.... 

My WPM

Honestly, I typed way faster when I was in high school playing games all day.
Here is my WPM by the way.

Lean Manufacturing / Improving facility

Okay... I have been trying to find an intern, part/full position during this summer for a company/facility to try my knowledge and skills in lean manufacturing and six sigmas learned at my school and previous works. Although I failed to do so, I am going to virtually share what I am going to do actually if I were to involved in actual project.

Since I was mainly hoping to work at a manufacturing facility, I will use simple manufacturing company as an example.

First, I would have to do some study and evaluate company's current situation. I have to gather data such as time, tasks assigned to each station, number of works, and all other necessary data.

Well, at this step, we need some theoretical standard numbers to compare those recorded data with to see if company is doing below or above the standard. These standards can be calculated using many different techniques such as

1. time study
2. motion study
3. queueing theory
4. assembly line balancing technique. 

If the company is doing below calculated standards, then management should step in and make some changes to maximize their utilization and reduce wastes.
To achieve this, company may need to redesign their facility more organized and efficient. They may also need to train their workers to do the job in proper way which maximizes their process. The most important thing is that the top management should be willing and able to make this changes even though it may cost them now.
At this step, company may incorporate such techniques as:

1. Facility redesign
2. Kaizen - gradual continuous improvement.
3. Poke-yoke - mostly done by organizing work stations and making somekind of signs.
4. PDCA/DMAIC - Plan Do Check Act/ Define Measure Analysis Improve Control. Both are for Six Sigma
5. Simulation(promodel) - to see if the theoretical changes actually improves the system (if used right, helps company save tons money)
6. 5s - Sort, Set in order, Shine, Standardize, Sustain

Once company made changes and improved their process, the most important thing is to sustain those changes by making check list. It would be nice thing to give rewards to department or people who made most improvements or suggested most realistic and helpful improvement to company to reinforce workers to be actively involved in improvement.