🗊Презентация Dynamic models and the Kalman filter

Категория: Физика
Нажмите для полного просмотра!
Dynamic models and the Kalman filter, слайд №1Dynamic models and the Kalman filter, слайд №2Dynamic models and the Kalman filter, слайд №3Dynamic models and the Kalman filter, слайд №4Dynamic models and the Kalman filter, слайд №5Dynamic models and the Kalman filter, слайд №6Dynamic models and the Kalman filter, слайд №7Dynamic models and the Kalman filter, слайд №8Dynamic models and the Kalman filter, слайд №9Dynamic models and the Kalman filter, слайд №10Dynamic models and the Kalman filter, слайд №11Dynamic models and the Kalman filter, слайд №12Dynamic models and the Kalman filter, слайд №13Dynamic models and the Kalman filter, слайд №14Dynamic models and the Kalman filter, слайд №15Dynamic models and the Kalman filter, слайд №16Dynamic models and the Kalman filter, слайд №17Dynamic models and the Kalman filter, слайд №18Dynamic models and the Kalman filter, слайд №19Dynamic models and the Kalman filter, слайд №20Dynamic models and the Kalman filter, слайд №21Dynamic models and the Kalman filter, слайд №22Dynamic models and the Kalman filter, слайд №23Dynamic models and the Kalman filter, слайд №24Dynamic models and the Kalman filter, слайд №25Dynamic models and the Kalman filter, слайд №26Dynamic models and the Kalman filter, слайд №27Dynamic models and the Kalman filter, слайд №28Dynamic models and the Kalman filter, слайд №29Dynamic models and the Kalman filter, слайд №30Dynamic models and the Kalman filter, слайд №31Dynamic models and the Kalman filter, слайд №32Dynamic models and the Kalman filter, слайд №33Dynamic models and the Kalman filter, слайд №34Dynamic models and the Kalman filter, слайд №35Dynamic models and the Kalman filter, слайд №36Dynamic models and the Kalman filter, слайд №37Dynamic models and the Kalman filter, слайд №38Dynamic models and the Kalman filter, слайд №39Dynamic models and the Kalman filter, слайд №40Dynamic models and the Kalman filter, слайд №41Dynamic models and the Kalman filter, слайд №42Dynamic models and the Kalman filter, слайд №43Dynamic models and the Kalman filter, слайд №44Dynamic models and the Kalman filter, слайд №45Dynamic models and the Kalman filter, слайд №46Dynamic models and the Kalman filter, слайд №47Dynamic models and the Kalman filter, слайд №48Dynamic models and the Kalman filter, слайд №49Dynamic models and the Kalman filter, слайд №50Dynamic models and the Kalman filter, слайд №51Dynamic models and the Kalman filter, слайд №52
Скачать

Содержание ▲

Вы можете ознакомиться и скачать презентацию на тему Dynamic models and the Kalman filter. Доклад-сообщение содержит 52 слайдов. Презентации для любого класса можно скачать бесплатно. Если материал и наш сайт презентаций Mypresentation Вам понравились – поделитесь им с друзьями с помощью социальных кнопок и добавьте в закладки в своем браузере.

Слайды и текст этой презентации

Слайд 1


Dynamic models and the Kalman filter, слайд №1
Описание слайда:

Слайд 2


Dynamic models and the Kalman filter, слайд №2
Описание слайда:

Слайд 3


Dynamic models and the Kalman filter, слайд №3
Описание слайда:

Слайд 4


Common Usage of These Techniques Macroeconomics, finance, time series models Autopilot, radar tracking Orbit tracking, satellite navigation (historically important) Speech, picture enhancement
Описание слайда:
Common Usage of These Techniques Macroeconomics, finance, time series models Autopilot, radar tracking Orbit tracking, satellite navigation (historically important) Speech, picture enhancement

Слайд 5


Another example Use nightlight data and the Kalman filter to adjust official GDP growth statistics. The idea is that economic activity is closely related to nightlight data. “Measuring Economic Growth from Outer Space” by Henderson, Storeygard, and Weil AER(2012)
Описание слайда:
Another example Use nightlight data and the Kalman filter to adjust official GDP growth statistics. The idea is that economic activity is closely related to nightlight data. “Measuring Economic Growth from Outer Space” by Henderson, Storeygard, and Weil AER(2012)

Слайд 6


Measuring Long-Term Growth
Описание слайда:
Measuring Long-Term Growth

Слайд 7


Measuring Short-Term Growth
Описание слайда:
Measuring Short-Term Growth

Слайд 8


Measuring Short-Term Growth
Описание слайда:
Measuring Short-Term Growth

Слайд 9


Dynamic models and the Kalman filter, слайд №9
Описание слайда:

Слайд 10


Dynamic models and the Kalman filter, слайд №10
Описание слайда:

Слайд 11


Dynamic models and the Kalman filter, слайд №11
Описание слайда:

Слайд 12


Dynamic models and the Kalman filter, слайд №12
Описание слайда:

Слайд 13


Dynamic models and the Kalman filter, слайд №13
Описание слайда:

Слайд 14


Dynamic models and the Kalman filter, слайд №14
Описание слайда:

Слайд 15


Dynamic models and the Kalman filter, слайд №15
Описание слайда:

Слайд 16


Basic Setup What if you know that are serially correlated: and , Then so one of the assumptions is violated! What to do? Can you still apply the model?
Описание слайда:
Basic Setup What if you know that are serially correlated: and , Then so one of the assumptions is violated! What to do? Can you still apply the model?

Слайд 17


Dynamic models and the Kalman filter, слайд №17
Описание слайда:

Слайд 18


Dynamic models and the Kalman filter, слайд №18
Описание слайда:

Слайд 19


Dynamic models and the Kalman filter, слайд №19
Описание слайда:

Слайд 20


Dynamic models and the Kalman filter, слайд №20
Описание слайда:

Слайд 21


Dynamic models and the Kalman filter, слайд №21
Описание слайда:

Слайд 22


Dynamic models and the Kalman filter, слайд №22
Описание слайда:

Слайд 23


Dynamic models and the Kalman filter, слайд №23
Описание слайда:

Слайд 24


Dynamic models and the Kalman filter, слайд №24
Описание слайда:

Слайд 25


Dynamic models and the Kalman filter, слайд №25
Описание слайда:

Слайд 26


Dynamic models and the Kalman filter, слайд №26
Описание слайда:

Слайд 27


Dynamic models and the Kalman filter, слайд №27
Описание слайда:

Слайд 28


Kalman Filter: Introduction State Space Representation [univariate case]: Notation: is the best linear predictor of st conditional on the information up to t-1. is the best linear predictor of yt conditional on the information up to t-1. is the best linear predictor of st conditional on the information up to t.
Описание слайда:
Kalman Filter: Introduction State Space Representation [univariate case]: Notation: is the best linear predictor of st conditional on the information up to t-1. is the best linear predictor of yt conditional on the information up to t-1. is the best linear predictor of st conditional on the information up to t.

Слайд 29


Kalman Filter: Main Idea Moving from t-1 to t Suppose we know and at time t-1. When arrive in period t we observe and Need to obtain st|t ! If we know , using the state equation: using the observation equation: yt+1|t = axt+1 + bst+1|t The key question: how to obtain st|t from ?
Описание слайда:
Kalman Filter: Main Idea Moving from t-1 to t Suppose we know and at time t-1. When arrive in period t we observe and Need to obtain st|t ! If we know , using the state equation: using the observation equation: yt+1|t = axt+1 + bst+1|t The key question: how to obtain st|t from ?

Слайд 30


Kalman Filter: Main Idea How to update st|t ? Idea: use the observed prediction error to infer the state at time t, It turns out it is optimal to update it using is called Kalman gain It measures how informative is the prediction error about the underlying state vector How do you think it depends on the variance of the observation error? It is chosen so that the new prediction error is orthogonal to all of the previous ones. Thus there is no (linear) predictable component in generated errors.
Описание слайда:
Kalman Filter: Main Idea How to update st|t ? Idea: use the observed prediction error to infer the state at time t, It turns out it is optimal to update it using is called Kalman gain It measures how informative is the prediction error about the underlying state vector How do you think it depends on the variance of the observation error? It is chosen so that the new prediction error is orthogonal to all of the previous ones. Thus there is no (linear) predictable component in generated errors.

Слайд 31


Kalman Filter: More Notations is the prediction error variance of given the history of observed variables up to t-1. is the prediction error variance of yt conditional on the information up to t-1. is the prediction error variance of conditional on the information up to t. Intuitively the Kalman gain is chosen so that is minimized. Will show this later.
Описание слайда:
Kalman Filter: More Notations is the prediction error variance of given the history of observed variables up to t-1. is the prediction error variance of yt conditional on the information up to t-1. is the prediction error variance of conditional on the information up to t. Intuitively the Kalman gain is chosen so that is minimized. Will show this later.

Слайд 32


Kalman Gain: Intuition Kalman gain is chosen so that is minimized. It can be shown that Intuition: If a big mistake is made forecasting ( is large), put a lot weight on the new observation (K is large). If the new information is noisy (R is large), put less weight on the new information (K is small).
Описание слайда:
Kalman Gain: Intuition Kalman gain is chosen so that is minimized. It can be shown that Intuition: If a big mistake is made forecasting ( is large), put a lot weight on the new observation (K is large). If the new information is noisy (R is large), put less weight on the new information (K is small).

Слайд 33


Kalman Filter: Example Kalman gain is Consider State equation Observation equation Additionally , where is a constant Assume that we picked (we don’t know anything about ). Can you calculate the Kalman gain in the 1st period, ? What is the interpretation?
Описание слайда:
Kalman Filter: Example Kalman gain is Consider State equation Observation equation Additionally , where is a constant Assume that we picked (we don’t know anything about ). Can you calculate the Kalman gain in the 1st period, ? What is the interpretation?

Слайд 34


Kalman Filter: The last step How do we get from to using ? Recall that for a bivariate normal distribution Using this property and the fact that Thus, st|t = st|t-1+bPt|t-1(Ft|t-1)-1(yt - yt|t-1) and Pt|t = Pt|t-1 – bPt|t-1(Ft|t-1)-1bPt|t-1
Описание слайда:
Kalman Filter: The last step How do we get from to using ? Recall that for a bivariate normal distribution Using this property and the fact that Thus, st|t = st|t-1+bPt|t-1(Ft|t-1)-1(yt - yt|t-1) and Pt|t = Pt|t-1 – bPt|t-1(Ft|t-1)-1bPt|t-1

Слайд 35


Kalman Filter: Finally From the previous slide st|t = st|t-1+bPt|t-1(Ft|t-1)-1(yt - yt|t-1) Pt|t = Pt|t-1 – bPt|t-1(Ft|t-1)-1bPt|t-1 Need: from to using Thus, we get the expression for the Kalman gain: Similarly And we are done!
Описание слайда:
Kalman Filter: Finally From the previous slide st|t = st|t-1+bPt|t-1(Ft|t-1)-1(yt - yt|t-1) Pt|t = Pt|t-1 – bPt|t-1(Ft|t-1)-1bPt|t-1 Need: from to using Thus, we get the expression for the Kalman gain: Similarly And we are done!

Слайд 36


Kalman Filter: Review We start from and . yt|t-1 = axt + bst|t-1 Calculate Kalman gain Update using observed Construct forecasts for the next period Repeat!
Описание слайда:
Kalman Filter: Review We start from and . yt|t-1 = axt + bst|t-1 Calculate Kalman gain Update using observed Construct forecasts for the next period Repeat!

Слайд 37


Kalman Filter: How to choose initial state If the sample size is large, the choice of the initial state is not very important In short samples can have significant effect For stationary models Where Solution to the last equation is Why? Under some very general conditions as
Описание слайда:
Kalman Filter: How to choose initial state If the sample size is large, the choice of the initial state is not very important In short samples can have significant effect For stationary models Where Solution to the last equation is Why? Under some very general conditions as

Слайд 38


Kalman Filter as a Recursive Regression Consider a regular regression function where Substituting From one of the previous slides: st|t = st|t-1+bPt|t-1(Ft|t-1)-1(yt - yt|t-1)
Описание слайда:
Kalman Filter as a Recursive Regression Consider a regular regression function where Substituting From one of the previous slides: st|t = st|t-1+bPt|t-1(Ft|t-1)-1(yt - yt|t-1)

Слайд 39


Kalman Filter as a Recursive Regression Consider a regular regression function where Substituting From one of the previous slides st|t = st|t-1+bPt|t-1(Ft|t-1)-1(yt - yt|t-1) Because
Описание слайда:
Kalman Filter as a Recursive Regression Consider a regular regression function where Substituting From one of the previous slides st|t = st|t-1+bPt|t-1(Ft|t-1)-1(yt - yt|t-1) Because

Слайд 40


Kalman Filter as a Recursive Regression Thus the Kalman filter can be interpreted as a recursive regression of a type where is the forecasting error at time t The Kalman filter describes how to recursively estimate
Описание слайда:
Kalman Filter as a Recursive Regression Thus the Kalman filter can be interpreted as a recursive regression of a type where is the forecasting error at time t The Kalman filter describes how to recursively estimate

Слайд 41


Optimality of the Kalman Filter Using the property of OLS estimates that constructed residuals are uncorrelated with regressors for all t Using the expression for and the state equation, it is easy to show that for all t and k=0..t-1 Thus the errors do not have any (linear) predictable component!
Описание слайда:
Optimality of the Kalman Filter Using the property of OLS estimates that constructed residuals are uncorrelated with regressors for all t Using the expression for and the state equation, it is easy to show that for all t and k=0..t-1 Thus the errors do not have any (linear) predictable component!

Слайд 42


Kalman Filter Some comments Within the class of linear (in observables) predictors the Kalman filter algorithm minimizes the mean squared prediction error (i.e., predictions of the state variables based on the Kalman filter are best linear unbiased): If the model disturbances are normally distributed, predictions based on the Kalman filter are optimal (its MSE is minimal) among all predictors: In this sense, the Kalman filter delivers optimal predictions.
Описание слайда:
Kalman Filter Some comments Within the class of linear (in observables) predictors the Kalman filter algorithm minimizes the mean squared prediction error (i.e., predictions of the state variables based on the Kalman filter are best linear unbiased): If the model disturbances are normally distributed, predictions based on the Kalman filter are optimal (its MSE is minimal) among all predictors: In this sense, the Kalman filter delivers optimal predictions.

Слайд 43


Dynamic models and the Kalman filter, слайд №43
Описание слайда:

Слайд 44


Kalman Filter Algorithm – Multivariate Case
Описание слайда:
Kalman Filter Algorithm – Multivariate Case

Слайд 45


Kalman Filter Algorithm – Multivariate Case (cont.)
Описание слайда:
Kalman Filter Algorithm – Multivariate Case (cont.)

Слайд 46


Kalman Filter Algorithm – Multivariate Case (cont.)
Описание слайда:
Kalman Filter Algorithm – Multivariate Case (cont.)

Слайд 47


Dynamic models and the Kalman filter, слайд №47
Описание слайда:

Слайд 48


Dynamic models and the Kalman filter, слайд №48
Описание слайда:

Слайд 49


Dynamic models and the Kalman filter, слайд №49
Описание слайда:

Слайд 50


Dynamic models and the Kalman filter, слайд №50
Описание слайда:

Слайд 51


Dynamic models and the Kalman filter, слайд №51
Описание слайда:

Слайд 52


Dynamic models and the Kalman filter, слайд №52
Описание слайда:

Скачать

Похожие презентации
Mypresentation.ru
Загрузить презентацию