EQ: How can we use linear models for non linear data?

Slides:



Advertisements
Similar presentations
4.1: Linearizing Data.
Advertisements

Chapter Four: More on Two- Variable Data 4.1: Transforming to Achieve Linearity 4.2: Relationships between Categorical Variables 4.3: Establishing Causation.
Lesson Nonlinear Regression: Transformations.
Transforming to Achieve Linearity
Section 12.2 Transforming to Achieve Linearity
+ Hw: pg 788: 37, 39, 41, Chapter 12: More About Regression Section 12.2b Transforming using Logarithms.
More about Relationships Between Two Variables
Chapter 4: More on Two-Variable (Bivariate) Data.
Scatter plots and Regression Algebra II. Linear Regression  Linear regression is the relationship between two variables when the equation is linear.
1 What you will learn today 1. New vocabulary 2. How to determine if data points are related 3. How to develop a linear regression equation 4. How to graph.
M ORE ON T WO V ARIABLE D ATA Non-Linear Data Get That Program Fishy Moths.
Chapter 11 Section 11.1 – 11.7 Review. Chapter 11.1 – 11.4 Pretest Evaluate each expression 1. (⅔) -4 = ___________2. (27) - ⅔ = __________ 3. (3x 2 y.
Cancer Cell Activity (p. 258). It will not always be obvious which transformation will work best; be willing to experiment. Hierarchy of Powers!
4.1 Modeling Nonlinear Data.  Create scatter plots of non linear data  Transform nonlinear data to use for prediction  Create residual plots.
Transforming Data.  P ,4  P ,7,9  Make a scatterplot of data  Note non-linear form  Think of a “common-sense” relationship.
Holt McDougal Algebra 2 Logarithmic Functions Holt Algebra 2Holt McDougal Algebra 2 How do we write equivalent forms for exponential and logarithmic functions?
Logarithms 1 Converting from Logarithmic Form to Exponential Form and Back 2 Solving Logarithmic Equations & Inequalities 3 Practice Problems.
SOLVING LOGARITHMIC EQUATIONS Objective: solve equations with a “log” in them using properties of logarithms How are log properties use to solve for unknown.
AP Statistics Section 4.1 A Transforming to Achieve Linearity.
Residual Plots Unit #8 - Statistics.
AP STATISTICS Section 4.1 Transforming to Achieve Linearity.
Solving Equations Using Addition or Subtraction Objective: Students will solve linear equations using addition and subtraction.
Exponential and log graphs Logarithms Identify all the graphs in the diagrams below. (2,16) (0,1) 0 y x  (3,125) 0 y x (0,1)  0 y 1 (4,16) x  0 y.
Chapter 10 Notes AP Statistics. Re-expressing Data We cannot use a linear model unless the relationship between the two variables is linear. If the relationship.
A little VOCAB.  Causation is the "causal relationship between conduct and result". That is to say that causation provides a means of connecting conduct.
Warm Ups Term 4 Week 5. Warm Ups 4/13/15 1.Solve the equation: log x = Find the inverse of the function: f(x) = 2 – x 3.
15.7 Curve Fitting. With statistical applications, exact relationships may not exist  Often an average relationship is used Regression Analysis: a collection.
Copyright © 2011 Pearson, Inc. 3.3 Logarithmic Functions and Their Graphs.
 Understand why re-expressing data is useful  Recognize when the pattern of the data indicates that no re- expression will improve it  Be able to reverse.
12.2 TRANSFORMING TO ACHIEVE LINEARITY To use transformations involving powers, roots, and logarithms to find a power or exponential model that describes.
Aim: What is the logarithms?
Chapter 12: More About Regression
Lesson 15-7 Curve Fitting Pg 842 #4 – 7, 9 – 12, 14, 17 – 26, 31, 38
WARM UP: Use the Reciprocal Model to predict the y for an x = 8
Section 3.4 Solving Exponential and Logarithmic Equations
A little VOCAB.
Solving Exponential Equations
Logarithmic Functions and Their Graphs
Chapter 12: More About Regression
Solving Exponential Equations
Bell Ringer Make a scatterplot for the following data.
Chapter 10 Re-Expressing data: Get it Straight
Ch. 12 More about regression
Equation Solving and Modeling
Chapter 12: More About Regression
Solving Linear Systems Algebraically
1. Find the inverse of the function y = 3x – 5.
Advanced Placement Statistics Section 4
y = logax y = ax a0 = 1 a1 = a y = axb y = abx
Residuals and Residual Plots
Solve for x: log3x– log3(x2 – 8) = log38x
Exponential and Logarithmic Functions
Chapter 12: More About Regression
Chapter 12: More About Regression
Ch 4 : More on Two-Variable Data
Aim: What is the logarithms?
Logarithmic Functions
4.1 Transformations.
Chapter 12: More About Regression
Chapter 12: More About Regression
Homework: pg. 276 #5, 6 5.) A. The relationship is strong, negative, and curved. The ratios are all Since the ratios are all the same, the exponential.
4 minutes Warm-Up Write each expression as a single logarithm. Then simplify, if possible. 1) log6 6 + log6 30 – log6 5 2) log6 5x + 3(log6 x – log6.
Chapter 12: More About Regression
Chapter 12: More About Regression
Transforming Relationships
Chapters Important Concepts and Terms
2.5 Correlation and Best-Fitting Lines
EQ: How can we use linear models for non linear data?
Scatterplots Regression, Residuals.
15.7 Curve Fitting.
Presentation transcript:

EQ: How can we use linear models for non linear data? Straightening Data

River water velocity and Distance from shore (cm/s) .5 22 1.5 23.18 2.5 25.48 3.5 25.25 4.5 27.15 5.5 27.83 6.5 28.49 7.5 28.18 8.5 28.50 9.5 28.63

Linear model analysis

Straighten the data

Writing the model Original model: Transformed model:

Using the model Predict the velocity of the river water at a distance of 5 meters from the shore

Why is a linear model not the best choice? Pattern, large residuals, no balance

Most popular transformations Exponential: Logarithmic: Power:

Exponential X Y 1 2 3 4 5 16 7 64 8 128 9 256 10 512

Straighten the data X Log(y) 1 2 0.301 3 0.602 4 0.903 5 1.204 7 1.505 2 0.301 3 0.602 4 0.903 5 1.204 7 1.505 8 1.806 9 2.107 10 2.709

Use the model Predict the y value for an x of 2.5

Find the model Mammal Weight (kg) Heart Rate (BPM) Mouse 0.03 580 At 0.32 320 Rabbit 3.97 170 Monkey 6.55 150 Dog 16 120 Elephant 2500 25

Power model Predict the heart rate for humans who way 60 kg.

Straightening the data The ONLY tools available for making models are linear ones. The only way to deal with non linear data is to make it linear Determine what type of model would be best to model the data. The data appears exponential Undo the operation that made the data The inverse of exponential functions are log functions

Transformed model Use the new model to predict for x = 5. Does it match the original data? Transformed model means transformed predictions.

How the exponential regression works Linear model: y = a + bx Exponential model: y = abx Determine that an exponential model is appropriate Transform the data Obtain a new linear model of the transformed data Transform the model back to the original units

Transforming the model New model: y = a + bx Which variable was transformed? y How was it transformed? Log(y) New model: log(y) = a + bx Solve for y in the new model

Algebra

Savings Accounts Sketch a graph of the original data Sketch a graph of the years and log(money) Is an exponential model appropriate? Sketch the residual plot to justify Transform the new linear model into an exponential model Predict for year 60

Harley-Davidson stock prices Sketch a graph of the original data Sketch a graph of the data and log(price) Is an exponential model appropriate? Sketch the residual plot to justify Transform the new linear model into an exponential model Predict for year 60

Other Models Open the file on heart beats Determine if a linear model is appropriate Determine if an exponential model is appropriate

Linear Model Exponential Model

Power Model Transforms both x and y by logarithms Take the log of the x’s and examine the graph for a linear relationship

Transforming the model New model: y = a + bx Which variable was transformed? y and x How was it transformed? Log(y) and log(x) New model: log(y) = a + blog(x) Eliminate the logs in the equation

Algebra

Airplane speed and flight lengths Sketch a graph of the original data Sketch a graph of the mph and log(miles) and examine the residuals. Sketch a graph of the log(mph) and log(miles) and examine the residuals. Determine which model is best Transform the new linear model into an exponential model.