MS&E 107 Lecture Notes - Lecture 9: Markov Chain, Standard Deviation, Statistical Hypothesis Testing
26 Oct 2016
School
Department
Course
Professor
Document Summary
Logistics: midterm: reschedule opportunity, markov chain: insert row and column, drag formulas over for mmult, friday: modeling sipmath introduction. 3 in a group, 3 minutes long, video upload youtube, random presentation selection. Forecasting: causal forecasting: regression, neural nets, time series: monitor prices, demand, over time. Forecasting - gives rise to flaw of averages. Fit a straight line to data that then predicts for any future x value the future y value. There is a standard deviation of this distribution - standard error - around where the prices will be (around the predicted point given by linear regression) How to fit line to data: minimize errors in some sense - minimize square of errors. Line: indicates sales = intercept + slope * year. For standard normal distribution, 95% of distribution lies between + and - 2 standard deviations. Suppose want normal with mean = x and sd = y from standard normal.
