Question

8. A particular area in a town suffers a high burglary rate. A sample of 100 streets is taken and in each of the sampled streets, a sample of six similar houses is taken. The table below shows the number of sampled houses, which have had burglaries during the last six months. 12 345 6 No. of houses burgled x | 0 No. of streets f 39 38 18 4 0 10 () (a) State any assumptions needed to justify the use of a binomial model for the number of houses per street which have been burgled during the last six months. (b) Derive the maximum likelihood estimator of p, the probability that a house of the type sampled has been burgled during the last six months. (c) Determine the probabilities for the binomial model using your estimate of p, and without doing a formal test, comment on the fit. (ii) An insurance company works on the basis that the probability of a house being burgled over a six-month period is 0.18. Carry out a test to investigate whether the binomial model with this value of p provides a good fit for the data

Answer 1

Sollution The expeimen Kthe chaat wkethe a henae has been ho Exdem o Cos of being be for loich Reu Combined i n the Eperiment ic deponden t of eech othey rmount each tri has two ertler house gut cone (Pailune mteru. s bumgda (suecsos) Rach, trial ts Or t 14is qutz imdt to commdy benmoti tial ther fore binomial mo de l ) et b d, the obebilrts .f a. heme beln, bugle4 tnc dne of heute. fx slut XN Bim (b. P) Sumpe. of 100 tws we w have a. frenueny drtbution of x- l Henee - 3 vil the mLE of 6 Thorfore Eft 0 35t 3A 1240 4C10 100 O.9 CS Scernedih.IS (-) (1-1s)-0 ty1. CamScanner

t-9 2-1) (- (xx) ( P(xx) Cxpeta fre (N P() bkenred foeauney 0.3728 37.26= 37 39 O. 3997 39.97 4 39 o. 1787 17 87 18 .26 O D42 $7 0.007 6.0004 0.3 0 O0 03 Cor be prebet Con be Ri eCx) Calculat b f)G.x4 3-Yo.7+( 1) Sinae tha Obsrved freauney te optical fa 0n aflesr ou Ome thal if can be wll fitted by Binemid d bution meam tia) tr thal we ont H 0.1 Ve uong R w 4 p valucO. 24710.0 do mot reflect H Hence ad 5. Level of sigmficet we co al cul hal +Rare fore bimom'md Scanned with Ro gibon Scdoc modal orth. sp. probabilty Dis la. wel-At.

RGui (32-bit) Windows Help View Misc Packages Edit File STOP t-c (39 , 38,18,4,0,1, 0) > > prob-NULL > tcalculation of expected frequencies > for (i in 1:7) + (7-i): prob [i]=choose (6, i-1)* (.18(i-1 ) )* ( 1-.18) + +} > #creating data frame for the given data >mat data.frame (n =0:6, obsFreq f, expFreq 100*prob) > #expected frequencies are less than 5..So the classes are merged > mat$groups <- cut (mat$n, breaks c(0:3,7), right FALSE, labels c ("0","1", "2","3 or more")) > mat expFreq 39 30.400667142 n obsFreq groups 1 0 38 40.039903066 18 1 1E6352 3 or more 5 4 1.058787936 0.092966746 3 or mo 6 5 re 7 6 0.003401222 3 or more > obsFreqGroup <-1 > expFregGroup <- tapply (matsexpFreq, mat$groups, aum) > data.frame (obs obsFreqGroup, exp expFreqGroup) pply (matsobsFreq, matsgroups, sum) obs exp C 39 30.400667 38 40.039903 2 18 21.973118 5 7.586312 3 or more > #chisq test for goodness of fit with p.18 > saveChiTest <-chisq.test (obsFreqGroup, p expFreqGroup/00) > saveChiTest Chi-squared test for given probabilities data: obsFreqGroup X-squared = 4.1365, df = 3, p-value 0.2471