Functions vs.
Data
Linear Regression
Linear Regression - Introduction
Bernd Schr oder
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Random Inuences When Evaluating Functions?
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Random Inuences When Evaluating Functions?
1. A function express a deterministic relation between an input variable x and an output variable y = f (x).
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Random Inuences When Evaluating Functions?
1. A function express a deterministic relation between an input variable x and an output variable y = f (x). 2. But if the input and/or output values of a function are measurements, then randomness creeps in.
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Random Inuences When Evaluating Functions?
1. A function express a deterministic relation between an input variable x and an output variable y = f (x). 2. But if the input and/or output values of a function are measurements, then randomness creeps in. 3. And at the rst stage of a scientic investigation, we do not know if there is a functional relationship between input and output or not.
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Random Inuences When Evaluating Functions?
1. A function express a deterministic relation between an input variable x and an output variable y = f (x). 2. But if the input and/or output values of a function are measurements, then randomness creeps in. 3. And at the rst stage of a scientic investigation, we do not know if there is a functional relationship between input and output or not. 4. If two variables are not deterministically related, then one does not inuence the other.
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Random Inuences When Evaluating Functions?
1. A function express a deterministic relation between an input variable x and an output variable y = f (x). 2. But if the input and/or output values of a function are measurements, then randomness creeps in. 3. And at the rst stage of a scientic investigation, we do not know if there is a functional relationship between input and output or not. 4. If two variables are not deterministically related, then one does not inuence the other. Paired measurements can look random.
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Random Inuences When Evaluating Functions?
1. A function express a deterministic relation between an input variable x and an output variable y = f (x). 2. But if the input and/or output values of a function are measurements, then randomness creeps in. 3. And at the rst stage of a scientic investigation, we do not know if there is a functional relationship between input and output or not. 4. If two variables are not deterministically related, then one does not inuence the other. Paired measurements can look random. 5. But there could be an underlying deterministic relation (a function) that is inuenced by measurement errors.
logo1 Louisiana Tech University, College of Engineering and Science
Bernd Schr oder Linear Regression - Introduction
Functions vs. Data
Linear Regression
Random Inuences When Evaluating Functions?
1. A function express a deterministic relation between an input variable x and an output variable y = f (x). 2. But if the input and/or output values of a function are measurements, then randomness creeps in. 3. And at the rst stage of a scientic investigation, we do not know if there is a functional relationship between input and output or not. 4. If two variables are not deterministically related, then one does not inuence the other. Paired measurements can look random. 5. But there could be an underlying deterministic relation (a function) that is inuenced by measurement errors. (Possibly most famous example: Keplers Laws.)
Bernd Schr oder Linear Regression - Introduction logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
When Measuring a Deterministic Relationship, Randomness Creeps In
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
When Measuring a Deterministic Relationship, Randomness Creeps In
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
When Measuring a Deterministic Relationship, Randomness Creeps In
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Simple Linear Regression Model
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Simple Linear Regression Model
1. Assumption: For any xed value of the independent variable x, the dependent variable is Y = 0 + 1 x + .
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Simple Linear Regression Model
1. Assumption: For any xed value of the independent variable x, the dependent variable is Y = 0 + 1 x + . 2. is assumed to be a normal random variable with expected value 0 and standard deviation .
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Simple Linear Regression Model
1. Assumption: For any xed value of the independent variable x, the dependent variable is Y = 0 + 1 x + . 2. is assumed to be a normal random variable with expected value 0 and standard deviation . 3. Y x := E(Y |x ) is the expected value of Y when x = x .
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Simple Linear Regression Model
1. Assumption: For any xed value of the independent variable x, the dependent variable is Y = 0 + 1 x + . 2. is assumed to be a normal random variable with expected value 0 and standard deviation . 3. Y x := E(Y |x ) is the expected value of Y when x = x . 2 4. Y x = V (Y |x ) is the variance of Y when x = x .
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Simple Linear Regression Model
1. Assumption: For any xed value of the independent variable x, the dependent variable is Y = 0 + 1 x + . 2. is assumed to be a normal random variable with expected value 0 and standard deviation . 3. Y x := E(Y |x ) is the expected value of Y when x = x . 2 4. Y x = V (Y |x ) is the variance of Y when x = x . 5. Clearly Y x = 0 + 1 x
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Simple Linear Regression Model
1. Assumption: For any xed value of the independent variable x, the dependent variable is Y = 0 + 1 x + . 2. is assumed to be a normal random variable with expected value 0 and standard deviation . 3. Y x := E(Y |x ) is the expected value of Y when x = x . 2 4. Y x = V (Y |x ) is the variance of Y when x = x . 2 2 5. Clearly Y x = 0 + 1 x , Y x =
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Quick Made-up Example to Illustrate the Model
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Quick Made-up Example to Illustrate the Model
On average, for the rst 30 days of training, a certain cross country runners mile time in seconds depends on the number of days d of training via the relationship T = 420 4d.
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Quick Made-up Example to Illustrate the Model
On average, for the rst 30 days of training, a certain cross country runners mile time in seconds depends on the number of days d of training via the relationship T = 420 4d. The actual time measured on a given day is skewed off this average prediction by a random distortion with mean zero and standard deviation 20 s.
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Quick Made-up Example to Illustrate the Model
On average, for the rst 30 days of training, a certain cross country runners mile time in seconds depends on the number of days d of training via the relationship T = 420 4d. The actual time measured on a given day is skewed off this average prediction by a random distortion with mean zero and standard deviation 20 s. The deterministic relationship gives the expected value for any given day.
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Quick Made-up Example to Illustrate the Model
On average, for the rst 30 days of training, a certain cross country runners mile time in seconds depends on the number of days d of training via the relationship T = 420 4d. The actual time measured on a given day is skewed off this average prediction by a random distortion with mean zero and standard deviation 20 s. The deterministic relationship gives the expected value for any given day. But the actual measurement on a given day need not be equal to this expected value.
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Quick Made-up Example to Illustrate the Model
On average, for the rst 30 days of training, a certain cross country runners mile time in seconds depends on the number of days d of training via the relationship T = 420 4d. The actual time measured on a given day is skewed off this average prediction by a random distortion with mean zero and standard deviation 20 s. The deterministic relationship gives the expected value for any given day. But the actual measurement on a given day need not be equal to this expected value. Inuences that can be considered random without further information (diet, sleep, mood, etc.) also affect the days performance.
logo1 Louisiana Tech University, College of Engineering and Science
Bernd Schr oder Linear Regression - Introduction
Functions vs. Data
Linear Regression
Idea for Linear Regression
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Idea for Linear Regression
In linear regression, we try to nd the (unknown) equation for quantities like T in the previous example from measured values over time.
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Idea for Linear Regression
In linear regression, we try to nd the (unknown) equation for quantities like T in the previous example from measured values over time. Note that if the relationship is not linear, linear regression will not be reliable.
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
Functions vs. Data
Linear Regression
Idea for Linear Regression
In linear regression, we try to nd the (unknown) equation for quantities like T in the previous example from measured values over time. Note that if the relationship is not linear, linear regression will not be reliable. For details, consider a course on regression analysis.
Bernd Schr oder Linear Regression - Introduction
logo1 Louisiana Tech University, College of Engineering and Science
