Class 10 tuesday oct 12
1 / 24

Class 10: Tuesday, Oct. 12.


79 views
Uploaded on:
Category: General / Misc
Description
Class 10: Tuesday, Oct. 12. Hurricane data set, review of confidence intervals and hypothesis tests Confidence intervals for mean response Prediction intervals Transformations Upcoming: Thursday: Finish transformations, Example Regression Analysis Tuesday: Review for midterm
Transcripts
Slide 1

Class 10: Tuesday, Oct. 12 Hurricane information set, survey of certainty interims and speculation tests Confidence interims for mean reaction Prediction interims Transformations Upcoming: Thursday: Finish changes, Example Regression Analysis Tuesday: Review for midterm Thursday: Midterm Fall Break!

Slide 2

Hurricane Data Is there a pattern in the quantity of sea tempests in the Atlantic after some time (perhaps an expand as a result of an Earth-wide temperature boost)? hurricane.JMP contains information on the quantity of storms in the Atlantic bowl from 1950-1997.

Slide 4

Inferences for Hurricane Data Residual plots and typical quantile plots show that presumptions of linearity, steady fluctuation and ordinariness in straightforward direct relapse model are sensible. 95% certainty interim for slant (change in mean tropical storms between year t and year t+1): (- 0.086,0.012) Hypothesis Test of invalid speculation that slant equivalents zero: test measurement = - 1.52, p-esteem =0.13. We acknowledge since p-esteem > 0.05. No confirmation of a pattern in storms from 1950-1997.

Slide 5

Scale for translating p-values: A substantial p-quality is not solid confirmation for H 0 , it just demonstrates that there is not solid proof against H 0.

Slide 6

Inference in Regression Confidence interims for incline Hypothesis test for slant Confidence interims for mean reaction Prediction interims

Slide 7

Car Price Example An utilized auto merchant needs to see how odometer perusing influences the offering cost of utilized autos. The merchant arbitrarily chooses 100 three-year old Ford Tauruses that were sold at closeout amid the previous month. Every auto was in top condition and furnished with programmed transmission, AM/FM tape cassette deck and aerating and cooling. carprices.JMP contains the value and number of miles on the odometer of every auto.

Slide 9

The utilized auto merchant has a chance to offer on a considerable measure of autos offered by a rental organization. The rental organization has 250 Ford Tauruses, all furnished with programmed transmission, ventilating and AM/FM tape cassette decks. The greater part of the autos in this parcel have around 40,000 miles on the odometer. The merchant would like an assessment of the normal offering cost of all autos of this sort with 40,000 miles on the odometer, i.e., E(Y|X=40,000). The slightest squares assessment is

Slide 10

Confidence Interval for Mean Response Confidence interim for E(Y|X=40,000): A scope of conceivable qualities for E(Y|X=40,000) taking into account the specimen. Inexact 95% Confidence interim: Notes about recipe for SE: Standard mistake gets to be littler as test size n builds, standard lapse is littler the closer is to In JMP, after Fit Line, click red triangle alongside Linear Fit and snap Confid Curves Fit. Utilize the crosshair apparatus by clicking Tools, Crosshair to locate the precise estimations of the certainty interim endpoints for a given X 0 .

Slide 12

A Prediction Problem The utilized auto merchant is offered a specific 3-year old Ford Taurus furnished with programmed transmission, aeration and cooling system and AM/FM tape cassette deck and with 40,000 miles on the odometer. The merchant might want to anticipate the offering cost of this specific auto. Best forecast taking into account slightest squares gauge:

Slide 13

Range of Selling Prices for Particular Car The merchant is keen on the scope of offering costs that this specific auto with 40,000 miles on it is prone to have. Under straightforward direct relapse model, Y|X takes after an ordinary circulation with mean and standard deviation . An auto with 40,000 miles on it will be in interim around 95% of the time. Class 5: We substituted the slightest squares gauges for and said auto with 40,000 miles on it will be in interim around 95% of the time. This is a decent close estimation however it disregards potential slip in minimum square gauges.

Slide 14

Prediction Interval 95% Prediction Interval: An interim that has give or take a 95% shot of containing the estimation of Y for a specific unit with X=X 0 ,where the specific unit is not in the first specimen. Rough 95% forecast interim: In JMP, after Fit Line, click red triangle by Linear Fit and snap Confid Curves Indiv. Utilize the crosshair apparatus by clicking Tools, Crosshair to locate the definite estimations of the forecast interim endpoints for a given X 0 .

Slide 16

A Violation of Linearity Y=Life Expectancy in 1999 X=Per Capita GDP (in US Dollars) in 1999 Data in gdplife.JMP Linearity presumption of basic straight relapse is plainly damaged. The increment in mean future for each extra dollar of GDP is less for vast GDPs than Small GDPs. Diminishing comes back to increments in GDP.

Slide 17

Transformations Violation of linearity: E(Y|X) is not a straight line. Changes: Perhaps E(f(Y)|g(X)) is a straight line, where f(Y) and g(X) are changes of Y and X, and a straightforward direct relapse model holds for the reaction variable f(Y) and informative variable g(X).

Slide 18

The mean of Life Expectancy | Log Per Capita has all the earmarks of being pretty nearly a straight line.

Slide 19

How would we utilize the change? Testing for relationship in the middle of Y and X: If the straightforward direct relapse model holds for f(Y) and g(X), then Y and X are related if and if the incline in the relapse of f(Y) and g(X) does not equivalent zero. P-esteem for test that incline is zero is <.0001: Strong confirmation that per capita GDP and future are related. Forecast and mean reaction: What might you anticipate the future to be for a nation with a for every capita GDP of $20,000?

Slide 20

How would we pick a change? Tukey’s Bulging Rule. See Handout. Match ebb and flow in information to the state of one of the bends attracted the four quadrants of the figure in the freebee. At that point utilize the related changes, selecting one for either X, Y or both.

Slide 21

Transformations in JMP Use Tukey’s Bulging standard (see present) to focus changes which may offer assistance. After Fit Y by X, click red triangle beside Bivariate Fit and snap Fit Special. Explore different avenues regarding changes recommended by Tukey’s Bulging tenet. Make leftover plots of the residuals for changed model versus the first X by clicking red triangle alongside Transformed Fit to … and clicking plot residuals. Pick changes which make the leftover plot have no example in the residuals' mean versus X. Analyze diverse changes by searching for change with littlest root mean square mistake on unique y-scale. In the event that utilizing a change that includes changing y, take a gander at root mean square lapse for fit measured on unique scale.

Slide 23

` By taking a gander at the root mean square mistake on the first y-scale, we see that the changes' majority enhance the untransformed model and that the change to log x is by a long shot the best.

Slide 24

The change to Log X seems to have for the most part evacuated a pattern in the residuals' mean. This implies that . There is still an issue of nonconstant change.

Slide 25

.