This document provides an introduction to mathematical modeling. It defines mathematical models as simplified representations of real-world entities that are characterized by variables, parameters, and functional forms. The modeling process involves gathering real-world data, simplifying it, developing a mathematical model, solving the model, and verifying its accuracy. Difference equations are used to model change over discrete time periods, while differential equations model continuous change. Examples are provided of modeling population growth and interacting species through systems of difference equations.
This document discusses recent developments in causal inference methods. It contains summaries of talks on several causal inference topics:
1. Miguel Hernan discusses the g-formula approach and inverse probability weighting for estimating causal effects under confounding.
2. Judith Lok discusses marginal structural models and g-estimation of structural nested models for longitudinal data, which allow controlling for time-varying confounding.
3. James Robins discusses single world intervention graphs for representing counterfactuals and the g-formula for estimating effects of dynamic treatment regimes.
4. Tyler VanderWeele discusses approaches for causal mediation analysis, including the difference method and natural direct and indirect effects.
We can define heteroscedasticity as the condition in which the variance of the error term or the residual term in a regression model varies. As you can see in the above diagram, in the case of homoscedasticity, the data points are equally scattered while in the case of heteroscedasticity, the data points are not equally scattered.
Two Conditions:
1] Known Variance
2] Unknown Variance
This document discusses key statistical concepts including random variables, probability distributions, expected value, variance, and correlation. It defines discrete and continuous random variables and explains how probability distributions assign probabilities to the possible values of a random variable. It also defines important metrics like expected value and variance, and how they are calculated for discrete and continuous random variables. The document concludes by explaining correlation, how the correlation coefficient measures the strength and direction of linear association between two variables, and how it is calculated.
We can define heteroscedasticity as the condition in which the variance of the error term or the residual term in a regression model varies. As you can see in the above diagram, in the case of homoscedasticity, the data points are equally scattered while in the case of heteroscedasticity, the data points are not equally scattered.
Two Conditions:
1] Known Variance
2] Unknown Variance
Data Science - Part XII - Ridge Regression, LASSO, and Elastic NetsDerek Kane
The document discusses various regression techniques including ridge regression, lasso regression, and elastic net regression. It begins with an overview of advancements in regression analysis since the late 1800s/early 1900s enabled by increased computing power. Modern high-dimensional data often has many independent variables, requiring improved regression methods. The document then provides technical explanations and formulas for ordinary least squares regression, ridge regression, lasso regression, and their properties such as bias-variance tradeoffs. It explains how ridge and lasso regression address limitations of OLS through regularization that shrinks coefficients.
This document provides an overview of correlation and linear regression. It defines key terms like independent variable, dependent variable, correlation coefficient, and regression coefficients. It explains how to calculate the correlation coefficient and regression coefficients using the least squares method. Properties of the regression coefficients are also discussed. Examples are provided to demonstrate how to interpret correlation, draw scatter plots, calculate coefficients, and predict values using the linear regression equation.
Stuck with your Regression Assignment? Get 24/7 help from tutors with Phd in the subject. Email us at support@helpwithassignment.com
Reach us at http://www.HelpWithAssignment.com
Intro to Quantitative Investment (Lecture 3 of 6)Adrian Aley
This document provides an overview of pairs trading and mean reversion trading strategies. It discusses how to construct trading pairs based on cointegration between assets, and outlines sample pairs trading strategies using techniques like distance methods and z-transforms. Key aspects covered include cointegration testing, vector error correction models, Granger causality, and the Engle-Granger two-step approach for cointegration analysis. The document is presented by Dr. Haksun Li and provides references and examples to illustrate concepts in pairs and mean reversion trading.
This document discusses recent developments in causal inference methods. It contains summaries of talks on several causal inference topics:
1. Miguel Hernan discusses the g-formula approach and inverse probability weighting for estimating causal effects under confounding.
2. Judith Lok discusses marginal structural models and g-estimation of structural nested models for longitudinal data, which allow controlling for time-varying confounding.
3. James Robins discusses single world intervention graphs for representing counterfactuals and the g-formula for estimating effects of dynamic treatment regimes.
4. Tyler VanderWeele discusses approaches for causal mediation analysis, including the difference method and natural direct and indirect effects.
We can define heteroscedasticity as the condition in which the variance of the error term or the residual term in a regression model varies. As you can see in the above diagram, in the case of homoscedasticity, the data points are equally scattered while in the case of heteroscedasticity, the data points are not equally scattered.
Two Conditions:
1] Known Variance
2] Unknown Variance
This document discusses key statistical concepts including random variables, probability distributions, expected value, variance, and correlation. It defines discrete and continuous random variables and explains how probability distributions assign probabilities to the possible values of a random variable. It also defines important metrics like expected value and variance, and how they are calculated for discrete and continuous random variables. The document concludes by explaining correlation, how the correlation coefficient measures the strength and direction of linear association between two variables, and how it is calculated.
We can define heteroscedasticity as the condition in which the variance of the error term or the residual term in a regression model varies. As you can see in the above diagram, in the case of homoscedasticity, the data points are equally scattered while in the case of heteroscedasticity, the data points are not equally scattered.
Two Conditions:
1] Known Variance
2] Unknown Variance
Data Science - Part XII - Ridge Regression, LASSO, and Elastic NetsDerek Kane
The document discusses various regression techniques including ridge regression, lasso regression, and elastic net regression. It begins with an overview of advancements in regression analysis since the late 1800s/early 1900s enabled by increased computing power. Modern high-dimensional data often has many independent variables, requiring improved regression methods. The document then provides technical explanations and formulas for ordinary least squares regression, ridge regression, lasso regression, and their properties such as bias-variance tradeoffs. It explains how ridge and lasso regression address limitations of OLS through regularization that shrinks coefficients.
This document provides an overview of correlation and linear regression. It defines key terms like independent variable, dependent variable, correlation coefficient, and regression coefficients. It explains how to calculate the correlation coefficient and regression coefficients using the least squares method. Properties of the regression coefficients are also discussed. Examples are provided to demonstrate how to interpret correlation, draw scatter plots, calculate coefficients, and predict values using the linear regression equation.
Stuck with your Regression Assignment? Get 24/7 help from tutors with Phd in the subject. Email us at support@helpwithassignment.com
Reach us at http://www.HelpWithAssignment.com
Intro to Quantitative Investment (Lecture 3 of 6)Adrian Aley
This document provides an overview of pairs trading and mean reversion trading strategies. It discusses how to construct trading pairs based on cointegration between assets, and outlines sample pairs trading strategies using techniques like distance methods and z-transforms. Key aspects covered include cointegration testing, vector error correction models, Granger causality, and the Engle-Granger two-step approach for cointegration analysis. The document is presented by Dr. Haksun Li and provides references and examples to illustrate concepts in pairs and mean reversion trading.
This document provides an overview of regularized regression techniques including ridge regression and lasso regression. It discusses when to use regularization to prevent overfitting, the tradeoff between bias and variance, and different types of regularization. Ridge regression minimizes the sum of squared coefficients while lasso regression minimizes the sum of absolute values of coefficients, allowing it to perform variable selection. Cross-validation is described as a method for selecting the optimal regularization parameter lambda. Advantages of regularization include improved generalization and interpretability. The document also provides an example using different regression models to predict diamond prices based on other variables in a dataset.
This document discusses multiple linear regression analysis. It begins by defining a multiple regression equation that describes the relationship between a response variable and two or more explanatory variables. It notes that multiple regression allows prediction of a response using more than one predictor variable. The document outlines key elements of multiple regression including visualization of relationships, statistical significance testing, and evaluating model fit. It provides examples of interpreting multiple regression output and using the technique to predict outcomes.
Intro to Quant Trading Strategies (Lecture 4 of 10)Adrian Aley
- The document introduces pairs trading via cointegration, where two assets that are cointegrated, or move together in the long run, can be traded to exploit short-term deviations from their long-term equilibrium.
- Cointegration means finding a linear combination of the two assets such that it is stationary. This stationary combination represents the long-run equilibrium between the assets.
- The document discusses testing for cointegration using augmented Dickey-Fuller tests, and outlines the vector error correction model (VECM) representation used to model cointegrated assets and implement pairs trading strategies.
linear regression is a linear approach for modelling a predictive relationship between a scalar response and one or more explanatory variables (also known as dependent and independent variables), which are measured without error. The case of one explanatory variable is called simple linear regression; for more than one, the process is called multiple linear regression. This term is distinct from multivariate linear regression, where multiple correlated dependent variables are predicted, rather than a single scalar variable. If the explanatory variables are measured with error then errors-in-variables models are required, also known as measurement error models.
In linear regression, the relationships are modeled using linear predictor functions whose unknown model parameters are estimated from the data. Such models are called linear models. Most commonly, the conditional mean of the response given the values of the explanatory variables (or predictors) is assumed to be an affine function of those values; less commonly, the conditional median or some other quantile is used. Like all forms of regression analysis, linear regression focuses on the conditional probability distribution of the response given the values of the predictors, rather than on the joint probability distribution of all of these variables, which is the domain of multivariate analysis.
Linear regression was the first type of regression analysis to be studied rigorously, and to be used extensively in practical applications.[4] This is because models which depend linearly on their unknown parameters are easier to fit than models which are non-linearly related to their parameters and because the statistical properties of the resulting estimators are easier to determine.
Linear regression has many practical uses. Most applications fall into one of the following two broad categories:
If the goal is error reduction in prediction or forecasting, linear regression can be used to fit a predictive model to an observed data set of values of the response and explanatory variables. After developing such a model, if additional values of the explanatory variables are collected without an accompanying response value, the fitted model can be used to make a prediction of the response.
If the goal is to explain variation in the response variable that can be attributed to variation in the explanatory variables, linear regression analysis can be applied to quantify the strength of the relationship between the response and the explanatory variables, and in particular to determine whether some explanatory variables may have no linear relationship with the response at all, or to identify which subsets of explanatory variables may contain redundant information about the response.
These are some slides I use in my Multivariate Statistics course to teach psychology graduate student the basics of structural equation modeling using the lavaan package in R. Topics are at an introductory level, for someone without prior experience with the topic.
This lecture discusses multiple regression analysis using two independent variables. Multiple regression faces two challenges: determining the influence of each variable and identifying which variables should be included. The model examines determinants of earnings using years of schooling and cognitive ability scores to predict earnings. Omitted variable bias can occur if an omitted variable is correlated with the included variables and influences earnings. Measures like R-squared, adjusted R-squared and F-tests evaluate how well the regression model fits the data.
Lecture 4 - Linear Regression, a lecture in subject module Statistical & Mach...Maninda Edirisooriya
Simplest Machine Learning algorithm or one of the most fundamental Statistical Learning technique is Linear Regression. This was one of the lectures of a full course I taught in University of Moratuwa, Sri Lanka on 2023 second half of the year.
This document provides guidance on performing and interpreting logistic regression analyses in SPSS. It discusses selecting appropriate statistical tests based on variable types and study objectives. It covers assumptions of logistic regression like linear relationships between predictors and the logit of the outcome. It also explains maximum likelihood estimation, interpreting coefficients, and evaluating model fit and accuracy. Guidelines are provided on reporting logistic regression results from SPSS outputs.
CHPTER 3: Multiple Linear Regression
Introduction
In simple regression we study the relationship between a dependent variable and a single explanatory (independent variable); assume that a dependent variable is influenced by only one explanatory variable.
Any business and economic applications of forecasting involve time series data. Re-gression models can be fit to monthly, quarterly, or yearly data using the techniques de-scribed in previous chapters. However, because data collected over time tend to exhibit trends, seasonal patterns, and so forth, observations in different time periods are re¬lated or autocorrelated. That is, for time series data, the sample of observations cannot be regarded as a random sample. Problems of interpretation can arise when standard regression methods are applied to observations that are related to one another over time. Fitting regression models to time series data must be done with considerable care.
This document summarizes seemingly unrelated regression (SURE) models. SURE models can handle multiple regression equations that may appear unrelated but are actually linked through correlated error terms. The key points are:
1) SURE models account for correlations between error terms in different regression equations, even if the equations do not share explanatory variables.
2) Estimation of SURE models involves generalized least squares to account for error term correlations between equations.
3) Estimating the error term covariance matrix Σ is important for SURE models, and can be done using restricted or unrestricted residuals from the separate equations.
Approximate Solution of a Linear Descriptor Dynamic Control System via a non-...IOSR Journals
This document discusses approximate solutions for linear descriptor dynamic control systems using a non-classical variational approach. It begins by introducing descriptor systems and their importance in applications. It then discusses making irregular systems regular through computational algorithms. The paper focuses on consistent initial conditions and their characterization. It proposes using a non-classical variational approach to obtain approximate solutions with a high degree of accuracy and freedom of choice for the bilinear form.
This document discusses strategies for designing factorial experiments with multiple factors. It explains that factorial experiments involve studying the effect of varying levels of factors on a response variable. The optimal design strategy depends on whether the circumstances are unusual or normal. For normal circumstances where there is some noise and factors influence each other, a fractional factorial or full factorial design is typically best. The document provides details on analyzing the data from factorial experiments to determine if factor effects and interactions are significant. It includes examples of calculating main effects and interactions from 2-level factorial data.
The document discusses mathematical modeling. It defines mathematical modeling as using mathematics to represent and analyze real-world phenomena. Mathematical models can be used to solve problems in fields like engineering, science, and economics. The document outlines the steps in the mathematical modeling process, including analyzing available data and governing principles, formulating models, and validating solutions. It also discusses different types of mathematical models, such as linear vs nonlinear, deterministic vs stochastic, static vs dynamic, discrete vs continuous, and quantitative vs qualitative models.
This document discusses supervised learning. Supervised learning uses labeled training data to train models to predict outputs for new data. Examples given include weather prediction apps, spam filters, and Netflix recommendations. Supervised learning algorithms are selected based on whether the target variable is categorical or continuous. Classification algorithms are used when the target is categorical while regression is used for continuous targets. Common regression algorithms discussed include linear regression, logistic regression, ridge regression, lasso regression, and elastic net. Metrics for evaluating supervised learning models include accuracy, R-squared, adjusted R-squared, mean squared error, and coefficients/p-values. The document also covers challenges like overfitting and regularization techniques to address it.
The document discusses decision trees and ensemble methods. It begins with an agenda that covers the bias-variance tradeoff, generalizations of this concept, the ExtraTrees algorithm, its sklearn interface, and conclusions. It then reviews decision trees, plotting sample data and walking through how the tree would split the data. Next, it covers the general CART algorithm and different impurity measures. It discusses controlling overfitting via tree depth and other techniques. Finally, it delves into explaining the bias-variance decomposition and tradeoff in more detail.
The document discusses the general linear model (GLM) as an extension of simple and multiple linear regression models. It describes how the GLM allows for modeling more complex relationships between variables by including transformed, squared, and interaction terms. Specifically, it explains how curvilinear relationships can be modeled by adding squared terms and how interaction effects between two variables can be modeled by including an interaction term. The document also discusses transforming the dependent variable to correct for non-constant variance.
The document provides an overview of regression analysis. It defines regression analysis as a technique used to estimate the relationship between a dependent variable and one or more independent variables. The key purposes of regression are to estimate relationships between variables, determine the effect of each independent variable on the dependent variable, and predict the dependent variable given values of the independent variables. The document also outlines the assumptions of the linear regression model, introduces simple and multiple regression, and describes methods for model building including variable selection procedures.
The document discusses linear inequalities in one variable. It defines a linear inequality in one variable as an inequality that can be written in the form ax + b > c, where a, b, and c are real numbers. It notes that the > symbol can be replaced by ≥, <, or ≤. The document provides examples and steps for transforming linear inequalities into standard form where the leading coefficient a is positive and the inequality is written as ax + b > 0. It emphasizes using properties of inequalities and multiplying by -1 when a is negative.
This document provides an overview of regularized regression techniques including ridge regression and lasso regression. It discusses when to use regularization to prevent overfitting, the tradeoff between bias and variance, and different types of regularization. Ridge regression minimizes the sum of squared coefficients while lasso regression minimizes the sum of absolute values of coefficients, allowing it to perform variable selection. Cross-validation is described as a method for selecting the optimal regularization parameter lambda. Advantages of regularization include improved generalization and interpretability. The document also provides an example using different regression models to predict diamond prices based on other variables in a dataset.
This document discusses multiple linear regression analysis. It begins by defining a multiple regression equation that describes the relationship between a response variable and two or more explanatory variables. It notes that multiple regression allows prediction of a response using more than one predictor variable. The document outlines key elements of multiple regression including visualization of relationships, statistical significance testing, and evaluating model fit. It provides examples of interpreting multiple regression output and using the technique to predict outcomes.
Intro to Quant Trading Strategies (Lecture 4 of 10)Adrian Aley
- The document introduces pairs trading via cointegration, where two assets that are cointegrated, or move together in the long run, can be traded to exploit short-term deviations from their long-term equilibrium.
- Cointegration means finding a linear combination of the two assets such that it is stationary. This stationary combination represents the long-run equilibrium between the assets.
- The document discusses testing for cointegration using augmented Dickey-Fuller tests, and outlines the vector error correction model (VECM) representation used to model cointegrated assets and implement pairs trading strategies.
linear regression is a linear approach for modelling a predictive relationship between a scalar response and one or more explanatory variables (also known as dependent and independent variables), which are measured without error. The case of one explanatory variable is called simple linear regression; for more than one, the process is called multiple linear regression. This term is distinct from multivariate linear regression, where multiple correlated dependent variables are predicted, rather than a single scalar variable. If the explanatory variables are measured with error then errors-in-variables models are required, also known as measurement error models.
In linear regression, the relationships are modeled using linear predictor functions whose unknown model parameters are estimated from the data. Such models are called linear models. Most commonly, the conditional mean of the response given the values of the explanatory variables (or predictors) is assumed to be an affine function of those values; less commonly, the conditional median or some other quantile is used. Like all forms of regression analysis, linear regression focuses on the conditional probability distribution of the response given the values of the predictors, rather than on the joint probability distribution of all of these variables, which is the domain of multivariate analysis.
Linear regression was the first type of regression analysis to be studied rigorously, and to be used extensively in practical applications.[4] This is because models which depend linearly on their unknown parameters are easier to fit than models which are non-linearly related to their parameters and because the statistical properties of the resulting estimators are easier to determine.
Linear regression has many practical uses. Most applications fall into one of the following two broad categories:
If the goal is error reduction in prediction or forecasting, linear regression can be used to fit a predictive model to an observed data set of values of the response and explanatory variables. After developing such a model, if additional values of the explanatory variables are collected without an accompanying response value, the fitted model can be used to make a prediction of the response.
If the goal is to explain variation in the response variable that can be attributed to variation in the explanatory variables, linear regression analysis can be applied to quantify the strength of the relationship between the response and the explanatory variables, and in particular to determine whether some explanatory variables may have no linear relationship with the response at all, or to identify which subsets of explanatory variables may contain redundant information about the response.
These are some slides I use in my Multivariate Statistics course to teach psychology graduate student the basics of structural equation modeling using the lavaan package in R. Topics are at an introductory level, for someone without prior experience with the topic.
This lecture discusses multiple regression analysis using two independent variables. Multiple regression faces two challenges: determining the influence of each variable and identifying which variables should be included. The model examines determinants of earnings using years of schooling and cognitive ability scores to predict earnings. Omitted variable bias can occur if an omitted variable is correlated with the included variables and influences earnings. Measures like R-squared, adjusted R-squared and F-tests evaluate how well the regression model fits the data.
Lecture 4 - Linear Regression, a lecture in subject module Statistical & Mach...Maninda Edirisooriya
Simplest Machine Learning algorithm or one of the most fundamental Statistical Learning technique is Linear Regression. This was one of the lectures of a full course I taught in University of Moratuwa, Sri Lanka on 2023 second half of the year.
This document provides guidance on performing and interpreting logistic regression analyses in SPSS. It discusses selecting appropriate statistical tests based on variable types and study objectives. It covers assumptions of logistic regression like linear relationships between predictors and the logit of the outcome. It also explains maximum likelihood estimation, interpreting coefficients, and evaluating model fit and accuracy. Guidelines are provided on reporting logistic regression results from SPSS outputs.
CHPTER 3: Multiple Linear Regression
Introduction
In simple regression we study the relationship between a dependent variable and a single explanatory (independent variable); assume that a dependent variable is influenced by only one explanatory variable.
Any business and economic applications of forecasting involve time series data. Re-gression models can be fit to monthly, quarterly, or yearly data using the techniques de-scribed in previous chapters. However, because data collected over time tend to exhibit trends, seasonal patterns, and so forth, observations in different time periods are re¬lated or autocorrelated. That is, for time series data, the sample of observations cannot be regarded as a random sample. Problems of interpretation can arise when standard regression methods are applied to observations that are related to one another over time. Fitting regression models to time series data must be done with considerable care.
This document summarizes seemingly unrelated regression (SURE) models. SURE models can handle multiple regression equations that may appear unrelated but are actually linked through correlated error terms. The key points are:
1) SURE models account for correlations between error terms in different regression equations, even if the equations do not share explanatory variables.
2) Estimation of SURE models involves generalized least squares to account for error term correlations between equations.
3) Estimating the error term covariance matrix Σ is important for SURE models, and can be done using restricted or unrestricted residuals from the separate equations.
Approximate Solution of a Linear Descriptor Dynamic Control System via a non-...IOSR Journals
This document discusses approximate solutions for linear descriptor dynamic control systems using a non-classical variational approach. It begins by introducing descriptor systems and their importance in applications. It then discusses making irregular systems regular through computational algorithms. The paper focuses on consistent initial conditions and their characterization. It proposes using a non-classical variational approach to obtain approximate solutions with a high degree of accuracy and freedom of choice for the bilinear form.
This document discusses strategies for designing factorial experiments with multiple factors. It explains that factorial experiments involve studying the effect of varying levels of factors on a response variable. The optimal design strategy depends on whether the circumstances are unusual or normal. For normal circumstances where there is some noise and factors influence each other, a fractional factorial or full factorial design is typically best. The document provides details on analyzing the data from factorial experiments to determine if factor effects and interactions are significant. It includes examples of calculating main effects and interactions from 2-level factorial data.
The document discusses mathematical modeling. It defines mathematical modeling as using mathematics to represent and analyze real-world phenomena. Mathematical models can be used to solve problems in fields like engineering, science, and economics. The document outlines the steps in the mathematical modeling process, including analyzing available data and governing principles, formulating models, and validating solutions. It also discusses different types of mathematical models, such as linear vs nonlinear, deterministic vs stochastic, static vs dynamic, discrete vs continuous, and quantitative vs qualitative models.
This document discusses supervised learning. Supervised learning uses labeled training data to train models to predict outputs for new data. Examples given include weather prediction apps, spam filters, and Netflix recommendations. Supervised learning algorithms are selected based on whether the target variable is categorical or continuous. Classification algorithms are used when the target is categorical while regression is used for continuous targets. Common regression algorithms discussed include linear regression, logistic regression, ridge regression, lasso regression, and elastic net. Metrics for evaluating supervised learning models include accuracy, R-squared, adjusted R-squared, mean squared error, and coefficients/p-values. The document also covers challenges like overfitting and regularization techniques to address it.
The document discusses decision trees and ensemble methods. It begins with an agenda that covers the bias-variance tradeoff, generalizations of this concept, the ExtraTrees algorithm, its sklearn interface, and conclusions. It then reviews decision trees, plotting sample data and walking through how the tree would split the data. Next, it covers the general CART algorithm and different impurity measures. It discusses controlling overfitting via tree depth and other techniques. Finally, it delves into explaining the bias-variance decomposition and tradeoff in more detail.
The document discusses the general linear model (GLM) as an extension of simple and multiple linear regression models. It describes how the GLM allows for modeling more complex relationships between variables by including transformed, squared, and interaction terms. Specifically, it explains how curvilinear relationships can be modeled by adding squared terms and how interaction effects between two variables can be modeled by including an interaction term. The document also discusses transforming the dependent variable to correct for non-constant variance.
The document provides an overview of regression analysis. It defines regression analysis as a technique used to estimate the relationship between a dependent variable and one or more independent variables. The key purposes of regression are to estimate relationships between variables, determine the effect of each independent variable on the dependent variable, and predict the dependent variable given values of the independent variables. The document also outlines the assumptions of the linear regression model, introduces simple and multiple regression, and describes methods for model building including variable selection procedures.
The document discusses linear inequalities in one variable. It defines a linear inequality in one variable as an inequality that can be written in the form ax + b > c, where a, b, and c are real numbers. It notes that the > symbol can be replaced by ≥, <, or ≤. The document provides examples and steps for transforming linear inequalities into standard form where the leading coefficient a is positive and the inequality is written as ax + b > 0. It emphasizes using properties of inequalities and multiplying by -1 when a is negative.
RPWORLD offers custom injection molding service to help customers develop products ramping up from prototypeing to end-use production. We can deliver your on-demand parts in as fast as 7 days.
2. WHAT IS A MATHEMATICAL MODEL?
1. Models are abstractions of reality!
2. Models are a representation of a particular thing, idea, or condition.
3. Mathematical Models are simplified representations of some real-world
entity
o Can be in equations or computer code
o Are intended to mimic essential features while leaving out
inessentials.
4. Mathematical Models are characterized by assumptions about:
o Variables (the things which change)
o Parameters (the things which do not change)
o Functional forms (the relationship between the two)
2
Introduction to Modeling - Introduction
3. MODELING PROCESS
3
Real World Data Model
Mathematical Conclusions
Predictions / Explanations
Simplifications
Interpretation
Reasonable?
yes – continue
no – go back to model
Verification – how good? Crunch numbers Solve equations
Introduction to Modeling - Introduction
4. MODELING CHANGE
Recall that the change can be modeled using the formula
change = future value − present value.
Change can take place over discrete or continuous time.
1. If the change occurs over discrete time periods (we know exactly what
comes next), then we get difference equations and we will actually be
observing the change.
2. If the change occurs continuously, then typically we will be observing the
rate of change and we will get a differential equation.
4
Introduction to Modeling - Modeling Change
5. FIRST DIFFERENCE EQUATIONS
DEFINITION
For a sequence of numbers 𝐴 = 𝑎0, 𝑎1, 𝑎2, … the first differences are
∆𝑎0 = 𝑎1 − 𝑎0
∆𝑎1 = 𝑎2 − 𝑎1
⋮
∆𝑎𝑛 = 𝑎𝑛+1 − 𝑎𝑛
Note: the first difference represents the rise or fall between
consecutive values of the sequence.
5
Introduction to Modeling - Modeling Change
6. EXERCISE 1 – DIFFERENCE EQUATIONS
Write a difference equation to represent the change during the 𝑛th
interval as a function of the previous term in the following sequences:
1. 1, 3, 7, 15, 31, …
2. 1, 2, 5, 11, 23, …
6
Introduction to Modeling - Modeling Change
7. DYNAMICAL SYSTEMS
Every difference equation
∆𝑎𝑛 = 𝑎𝑛+1 − 𝑎𝑛
determines a dynamical systems (relationship between terms of a sequence)
by solving for 𝑎𝑛+1 for a given initial value 𝑎0:
𝑎𝑛+1 = 𝑎𝑛 + ∆𝑎𝑛
A numerical solution is a table of values, 𝑎𝑗 for all 𝑗, satisfying the dynamical
system.
7
Introduction to Modeling - Modeling Change
8. EXERCISE 2 – SIMPLE MODEL
Formulate a dynamical system that
models the following situation:
You owe $1000 on a credit card after
buying Christmas gifts for your family
that charges 1.5% interest each
month. You pay $50 each month and
you make no new charges.
8
Introduction to Modeling - Modeling Change
9. SIMPLIFICATION OF OUR MODEL
One very powerful simplifying relationship is proportionality.
DEFINITION
Two variables 𝑦 and 𝑥 are proportional (to each other) if one is always
a constant multiple of the other. That is, if
𝑦 = 𝑘𝑥
for some nonzero constant 𝑘. We write 𝑦 ∝ 𝑥.
9
Introduction to Modeling - Approximation of Change
10. EXERCISE 3 – POPULATION GROWTH
The following data was collected
from an experiment measuring
the growth of a yeast culture.
Formulate a dynamical system
that models the growth of the
yeast culture.
10
Time in hours 𝒏 Observed yeast biomass 𝒑𝒏
0 9.6
1 18.3
2 29.0
3 47.2
4 71.1
5 119.1
6 174.6
7 257.3
Introduction to Modeling - Approximation of Change
To simplify the model, we assume that the change in the population grows
proportionally with respect to the current size of the population.
∆𝑝𝑛 = 𝑘𝑝𝑛 for some growth constant 𝑘
11. SUMMARY CHART
11
We see that our model is not a great fit for the data. This may be due to the
limited amount of original data
Introduction to Modeling - Approximation of Change
0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8
Biomass
Time in hours
Comparison Chart - Original data vs Projected Data
Original Data Projected Data
12. ISSUES WITH CONSTANT GROWTH RATE
If we assume the change grows proportionally with the current
population size,
∆𝑎𝑛 = 𝑘𝑎𝑛
then 𝑎𝑛+1 would grow without bound.
This may, or may not, be a good simplification. To relax this
simplification, we could assume the population has a limiting value, say
𝑀. Then we could define the growth rate 𝑘 by a limiting factor.
𝑘 = 𝑟 𝑀 − 𝑎𝑛 → ∆𝑎𝑛 = 𝑟 𝑀 − 𝑎𝑛 𝑎𝑛
12
Introduction to Modeling - Approximation of Change
13. EXERCISE 4 – CARRYING CAPACITY
The following data was
collected from an
experiment measuring the
growth of a yeast culture.
Formulate a dynamical
system that models the
growth of the yeast culture.
13
Time in
hours 𝒏
Observed yeast
biomass 𝒑𝒏
Time in
hours 𝒏
Observed yeast
biomass 𝒑𝒏
0 9.6 10 513.3
1 18.3 11 559.7
2 29.0 12 594.8
3 47.2 13 629.4
4 71.1 14 640.8
5 119.1 15 651.1
6 174.6 16 655.9
7 257.3 17 659.6
8 350.7 18 661.8
9 441.0
The population appears to
be approaching a limiting
value, or carrying capacity
Introduction to Modeling - Approximation of Change
14. SUMMARY CHART
14
With the addition of the limiting value, we see that our model is a better fit to
the original data Introduction to Modeling - Approximation of Change
0
100
200
300
400
500
600
700
0 2 4 6 8 10 12 14 16 18 20
Biomass
Time in hours
Comparison Chart - Original data vs Projected Data
Original data Projected Data
15. SOLUTIONS TO DYNAMICAL SYSTEMS
When we develop a mathematical model, we would like to be able to
find the solution; we would like an easy way to express
𝑎𝑛+1.
An obvious method for finding the solution of a simple model is to
hypothesize the form of a solution and then test if it is correct. This
method is known as the Method of Conjecture.
15
Introduction to Modeling - Analytical Solutions
16. THE METHOD OF CONJECTURE
The following summarize the procedure for applying this method.
1. Observe a pattern.
2. Conjecture a form of the solution to the dynamical system.
3. Test the conjecture by substitution.
4. Accept or reject the conjecture depending on whether it does or
does not satisfy the system after the substitution and algebraic
manipulation.
16
Introduction to Modeling - Analytical Solutions
17. LINEAR DYNAMICAL SYSTEMS
What if ∆𝑎𝑛 = 𝑐𝑎𝑛, where 𝑐 is some nonzero constant?
THEOREM
The solution of the linear dynamical system 𝑎𝑛+1 = 𝑟𝑎𝑛 for 𝑟 any
nonzero constant is
𝑎𝑘 = 𝑟𝑘
𝑎0
where 𝑎0 is a given initial value.
17
Introduction to Modeling - Analytical Solutions
18. EXERCISE 5 – SOLUTIONS
Find the solution to the difference equations in the following problems:
1. 𝑎𝑛+1 = 5𝑎𝑛, 𝑎0 = 10
2. 𝑎𝑛+1 =
3𝑎𝑛
4
, 𝑎0 = 64
18
Introduction to Modeling - Analytical Solutions
19. EQUILIBRIUM VALUE
DEFINITION
A number 𝑎 is called an equilibrium value or fixed point of a dynamical
system
𝑎𝑛+1 = 𝑓 𝑎𝑛
if 𝑓(𝑎𝑛) = 𝑎 for all 𝑛 ≥ 𝑁.
If 𝑓 𝑎𝑛 = 𝑎 for all 𝑛 ≥ 0, 𝑎 is a constant solution to the dynamical
system.
19
Introduction to Modeling - Analytical Solutions
20. LONG-TERM BEHAVIOR
Let’s consider some significant values of 𝑟.
20
𝑟 = 0 Constant solution and equilibrium value at 0
𝑟 = 1 All initial values are constant solutions and equilibrium
values.
𝑟 < 0 Oscillation
𝑟 < 1 Decay to limiting value of 0
𝑟 > 1 Growth without bound
Introduction to Modeling - Analytical Solutions
21. LINEAR DYNAMICAL SYSTEMS
What if ∆𝑎𝑛 = 𝑐𝑎𝑛 + 𝑏, where 𝑐 is some nonzero constant?
THEOREM
The equilibrium value for the dynamical system
𝑎𝑛+1 = 𝑟𝑎𝑛 + 𝑏, 𝑟 ≠ 1
is
𝑎 =
𝑏
1 − 𝑟
If 𝑟 = 1 and 𝑏 = 0, every number is an equilibrium value. If 𝑟 = 1 and
𝑏 ≠ 0, no equilibrium value exists.
21
Introduction to Modeling - Analytical Solutions
22. TYPES OF EQUILIBRIUMS
There are two kinds of equilibrium, called stable and unstable.
1. If the equilibrium is stable, then 𝑎𝑛 → 𝑎 as 𝑛 → ∞ regardless of 𝑎0.
2. If the equilibrium is unstable, then the equilibrium is achieved only
if 𝑎0 = 𝑎.
For the dynamical system 𝑎𝑛+1 = 𝑟𝑎𝑛 + 𝑏,
1. If 𝑟 < 1, the equilibrium is stable.
2. If 𝑟 > 1, the equilibrium is unstable.
22
Introduction to Modeling - Analytical Solutions
23. EXERCISE 6 - EQUILIBRIUMS
Suppose we give a patient a daily dosage of 0.2mg of a drug at the end
of the day. 40% of the drug remains in the patient’s system.
1. How much drug is in the patient’s system at time 𝑛 + 1?
2. What is the equilibrium value?
3. Examine what happens for 𝑎0 = 0.25, 𝑎0 = 0.30, and 𝑎0 = 0.35.
23
Introduction to Modeling - Analytical Solutions
24. LINEAR DYNAMICAL SYSTEMS
THEOREM
The solution of the dynamical system
𝑎𝑛+1 = 𝑟𝑎𝑛 + 𝑏, 𝑟 ≠ 1
is
𝑎𝑘 = 𝑟𝑘
𝑐 +
𝑏
1 − 𝑟
for some constant 𝑐 (which depends on the initial condition).
24
Introduction to Modeling - Analytical Solutions
25. SYSTEMS OF DIFFERENCE EQUATIONS
For the systems considered here, we find the equilibrium values and
then explore starting values in their vicinity. If we start close to an
equilibrium value, we want to know whether the system will:
1. Remain close
2. Approach the equilibrium value
3. Not remain close.
Does the long-term behavior described by the numerical solution
appear to be sensitive to the initial condition?
25
Introduction to Modeling - Systems
26. MODELS FOR INTERACTING POPULATIONS
When species interact the population dynamics of each species is
affected. We consider here two-species systems.
There are three main types of interaction.
1. If the growth rate of one population is decreased and the other
increased the populations are in a predator-prey situation.
2. If the growth rate of each population is decreased then it is
competition.
3. If each population’s growth rate is enhanced then it is called
symbiosis.
26
Introduction to Modeling - Systems
27. PREDATOR-PREY MODEL
If 𝑁𝑛 is the prey population and 𝑃𝑛 that of the predator at time 𝑛 then our
model is
∆𝑁𝑛 = 𝑁𝑛 𝑎 − 𝑏𝑃𝑛
∆𝑃𝑛 = 𝑃𝑛 𝑐𝑁𝑛 − 𝑑
where 𝑎, 𝑏, 𝑐 and 𝑑 are positive constants.
1. The prey in the absence of any predation grows unboundedly.
2. The effect of the predation is to reduce the prey’s per capita growth rate
by a term proportional to the prey and predators population.
3. In the absence of any prey for sustenance the predator’s death rate
results in exponential decay.
4. The prey’s contribution to the predator’s growth is proportional to the
available prey as well as to the size of the predator population.
27
Introduction to Modeling - Systems
28. EQUILIBRIUM VALUES
The equilibrium values for the system are those values for which no
change in the system takes place.
Let 𝑁 and 𝑃 be those values. Substitution in our model yields the
following requirements for the equilibrium values:
28
Introduction to Modeling - Systems
∆𝑁𝑛 = 0 = 𝑁 𝑎 − 𝑏𝑃
∆𝑃𝑛 = 0 = 𝑃 𝑐𝑁 − 𝑑
𝑁 = 0 or 𝑃 =
𝑎
𝑏
𝑃 = 0 or 𝑁 =
𝑑
𝑐
𝑃, 𝑁 → 0,0 𝑎𝑛𝑑
𝑎
𝑏
,
𝑑
𝑐
29. EXERCISE 7
Suppose we have Owls and Mice in a park and the Owls feed on the
Mice. Let 𝑀𝑛 and 𝑂𝑛 be the population at time 𝑛 and let’s say
𝑀𝑛+1 = 1.5𝑀𝑛 − 0.005𝑂𝑛𝑀𝑛
𝑂𝑛+1 = 0.5𝑂𝑛 + 0.001𝑂𝑛𝑀𝑛
1. What do these equations represent?
2. What are the equilibrium values?
3. What happens if 𝑀0 = 200 and 𝑂0 = 100?
29
Introduction to Modeling - Systems
30. SUMMARY
After the first 5 iterations it appears the mice population is growing while the
owl population increases.
After the 5th iteration, the owl population begins to grow but so does the
mice population.
At the 11th iteration, we see that the mice population has grown large
enough to allow the owl population to make large growth jumps which
allows the owls to take over.
Since the owls no longer have a food source, they begin to die off.
30
Introduction to Modeling - Systems
31. COMPETITIVE EXCLUSION
Here two or more species compete for the same limited food source or
in some way inhibit each other’s growth.
Let 𝐴𝑛 and 𝐵𝑛 be the populations of species 𝐴 and 𝐵 respectively at
time 𝑛 and suppose that in the absence of the other species, each
individual species exhibits unconstrained growth. Let 𝑘1 and 𝑘2 be the
positive growth rates.
∆𝐴𝑛 = 𝐴𝑛 𝑘1 − 𝑘3𝐵𝑛
∆𝐵𝑛 = 𝐵𝑛 𝑘2 − 𝑘4𝐴𝑛
31
Introduction to Modeling - Systems
32. COMPETITIVE EXCLUSION
Solving each equation for the 𝑛 + 1 term gives
𝐴𝑛+1 = 1 + 𝑘1 𝐴𝑛 − 𝑘3𝐵𝑛𝐴𝑛
𝐵𝑛+1 = 1 + 𝑘2 𝐵𝑛 − 𝑘4𝐴𝑛𝐵𝑛
When there is no change we have equilibrium values at
∆𝐴𝑛 = 0 = 𝐴𝑛 𝑘1 − 𝑘3𝐵𝑛
∆𝐵𝑛 = 0 = 𝐵𝑛 𝑘2 − 𝑘4𝐴𝑛
𝐴, 𝐵 = (0,0) and 𝐴, 𝐵 =
𝑘2
𝑘4
,
𝑘1
𝑘3
32
Introduction to Modeling - Systems
33. EXERCISE 8
Suppose a species of spotted owls competes for survival in a habitat that also
supports hawks. Let 𝑂𝑛 and 𝐻𝑛 be the population at time 𝑛 and let’s say
𝑂𝑛+1 = 1.2𝑂𝑛 − 0.001𝑂𝑛𝐻𝑛
𝐻𝑛+1 = 1.3𝐻𝑛 − 0.002𝑂𝑛𝐻𝑛
1. What do these equations represent?
2. What are the equilibrium values?
3. What happens if 𝑂0 = 151 and 𝐻0 = 199? What happens if 𝑂0 = 149
and 𝐻0 = 201? What happens if 𝑂0 = 10 and 𝐻0 = 10?
33
Introduction to Modeling - Systems
34. SYMBIOSIS
There are many examples where the interaction of two or more species
is to the advantage of all. Symbiosis often plays the crucial role in
promoting and even maintaining such species: plant and seed dispersal
is one example.
Let 𝐴𝑛 and 𝐵𝑛 be the populations of species 𝐴 and 𝐵 respectively at
time 𝑛. Let 𝑘1 and 𝑘2 be the positive growth rates.
∆𝐴𝑛 = 𝐴𝑛 𝑘1 + 𝑘3𝐵𝑛
∆𝐵𝑛 = 𝐵𝑛 𝑘2 + 𝑘4𝐴𝑛
34
Introduction to Modeling - Systems
35. THE MODELING PROCESS
Suppose we want to understand some behavior or phenomenon in the
real world. We may wish to make predictions about that behavior in
the future and analyze the effects that various situations have on it.
35
Mathematical World
1. Models
2. Equations/Formulas
3. Operations/Rules
4. Conclusions
Real World Systems
1. Observed Behavior
2. Phenomenon
Introduction to Modeling - Modeling Process
36. SYSTEM
A system is a collection of objects, typically represented by the
variables in the model, that interact or are interdependent.
We are interested in understanding how a particular systems works,
what causes changes in the system, and how sensitive the system is to
certain changes to help us predict what changes might occur and when
they occur.
EXAMPLE:
In our previous examples, what happened then we saw too many mice
in our system?
36
Introduction to Modeling - Modeling Process
37. REAL-WORLD VS MATH WORLD
37
Real-World System Mathematical Model
Mathematical
Conclusions
Real-World
Conclusions
Observation
Simplification
Analysis
Trials/Experiments
Interpretation
Introduction to Modeling - Modeling Process
38. WHY MODEL?
Why not just test things in the real world?
There are many situations in which we would not want to follow such a
course of action (even with the risk of loss of fidelity).
1. Level of concentration of a drug (lethal).
2. Radiation effects of a failure in a nuclear power plant.
3. Atmospheric changes
May not even be possible to produce a trial, not just for moral reasons.
38
Introduction to Modeling - Modeling Process
39. MATHEMATICAL MODEL
A mathematical model is defined as a mathematical construct designed
to study a particular real-world system or phenomenon. It can include
graphical, symbolic, simulation, and experimental constructs.
When developing our model, we should keep a few things in mind:
1. Feasibility: is the real-world system just to much?
2. Fidelity: how close does the model match reality?
3. Costs: do we have the resources, etc.?
4. Flexibility: can the model be expanded easily?
39
Introduction to Modeling - Modeling Process
40. CONSTRUCTION OF MODELS
The process of constructing a model is similar to the scientific method:
1. Identify the problem. This is not as straightforward as it sounds. Be
as clear as you can.
2. Make assumptions. Simplify relationships, reduce the number of
variables, classify the variables (independent, dependent, neither),
determine the relationships between the variables.
3. Solve or interpret the model. Solve the equations (possible
numerically).
40
Introduction to Modeling - Modeling Process
41. CONSTRUCTION OF MODELS
4. Verify the model. Did we answer the question, is it feasible, does it
make sense, compare with observations.
5. Implement the model. Make it useful/practical.
6. Maintain the model. Update as necessary.
NOTE:
This is like the scientific method for testing hypothesis and we do try to
be as scientific as possible, but we must remember that this is a
creative/artistic/intuitive part of modelling and that for the most part,
our models can never duplicate reality exactly.
41
Introduction to Modeling - Modeling Process
42. SIMPLIFYING OR REFINING
Model Simplification
1. Restrict problem identification.
2. Neglect variables.
3. Group effects of several
variables.
4. Set some variables to be
constant.
5. Assume simple (linear)
relationships.
6. Incorporate more assumptions.
Model Refinement
1. Expand the problem.
2. Consider additional variables.
3. Consider each variable in detail.
4. Allow variation in the variables.
5. Consider nonlinear
relationships.
6. Reduce the number of
assumptions.
42
Introduction to Modeling - Modeling Process
43. PROPERTIES OF A MODEL
The following are terms that are useful in describing models:
1. A model is said to be robust when its conclusions do not depend on
the precise satisfaction of the assumptions.
2. A model is fragile if its conclusions do depend on the precise
satisfaction of some sort of conditions.
3. Sensitivity refers to the degree of change in a model’s conclusions
as some condition on which they depend is varied; the greater the
change the more sensitive is the model to that condition.
43
Introduction to Modeling - Modeling Process