how to generate pdf in asp net mvc : Converting image pdf to text SDK software service wpf winforms asp.net dnn movieforecast2-part1749

office.
Among ordinal covariates the three significant variables were average user ratings (BAVG),
averagecriticratings(CRAVG)andgenderentropy(GENTR).Observe thatthestandardizedco-
efficientofBAVG (0.39)isalmosttwice aslargeasthestandardizedcoefficientofCRAVG (0.20).
Thisindicatesthataverageuserratingsaremoreinfluentialthanaverageprofessionalcriticreviews
inpredictinga movie’srevenue trajectory. Giventhe amountof attentionthatcritic ratingshave
receivedin the past, this resulthas considerable practical consequences. On the otherhand, the
simultaneous significance of BAVG and CRAVG,together with the relatively low correlation be-
tweenuserandcriticratings(Table3),reinforcesourearlierremarkthatthesetwovariablesshould
be consideredascomplementaryproxiesofa movie’srevenue potential. Finally,the significanceof
GENTR statesthatdiversityofamovie’sonlineraters(withrespecttogender)ispositivelycorre-
latedwithfuture revenues. ThisfindingisconsistentwithGodesandMayzlin’s(2004)resultthat
higherdispersionofword-of-mouthamongdifferentgroupscorrelateswithhigherfutureviewership
inthecontextof TVshows.
Interestingly,covariates relatingto marketing andearlybox-office revenueswere significantin
predicting coefficient P,but not inpredicting coefficientQ. Similarly, covariatesrelating to user
andcritic ratingswere significantinpredicting Q,butnotinpredictingP.Thisisconsistentwith
the theoretical interpretation of coefficients P and Q of equation (4) as capturing the intensity
of publicity and word-of-mouth respectively and reinforces the validity of using a modified Bass
equationtomodelthe evolutionof motionpicture revenues.
Observe, finally, thatthevolume ofratings(TOT)wasnotsignificantin predictingcoefficient
Q. Thisresultisnotsurprising, anddoesnot contradicttheresultsof Liu(2004)orDuanet. al.
(2005),bothofwhomfoundthevolumeofonlineconversationstobehighlysignificant. Toseethis,
observe that the structure of equation (4)already assumes thatthe impactof word-of-mouth on
revenuesis the product of coefficient Q multiplied by the (discounted) number of past adopters.
Ourdataindicatesthatthevolumeofratingsishighlycorrelatedwiththevolumeof sales(Figure
2). One therefore expects that, if equation (4) provides a correct description of the underlying
phenomenon,theimpactofthevolumeofratingswouldbeabsorbedbytheterm
Rt
¿=0
_
R(t¡¿)"
¿
d¿
andwouldnotbesignificantinpredictingcoefficientQ.Thefactthatthevolumeofratingswasnot
asignificantpredictorofQ,thus,constitutesafurtherconfirmation ofourmodelingassumptions.
21
Converting image pdf to text - SDK software service:C# PDF Convert to Text SDK: Convert PDF to txt files in C#.net, ASP.NET MVC, WinForms, WPF application
C# PDF to Text (TXT) Converting Library to Convert PDF to Text
www.rasteredge.com
Converting image pdf to text - SDK software service:VB.NET PDF Convert to Text SDK: Convert PDF to txt files in vb.net, ASP.NET MVC, WinForms, WPF application
VB.NET Guide and Sample Codes to Convert PDF to Text in .NET Project
www.rasteredge.com
epsilon 
0.1 
0.2 
0.3 
0.4 
0.5 
0.6 
0.7 
0.8 
0.9 
1.0 
0.1 
0.180 
1.960  4.986  5.756  6.292  5.570  2.440  1.021  1.833  1.971 
0.2 
0.242 
0.165 0.443 0.940 8.474 6.108 2.503 1.141 2.260 1.394 
0.3 
0.271 
0.213 0.156  0.271  0.551  6.536  5.403  2.275  2.331  3.160 
0.4 
0.465 
0.305  0.221 
0.150 
0.299  2.746  5.037  2.235  0.985  1.055 
0.5 
0.903 
0.578  0.395  0.260 0.144  0.382  4.978  2.978  0.969  1.163 
0.6 
1.584 
1.076  0.776  0.543  0.325 0.141  0.479  1.292  1.133  1.009 
0.7 
2.597 
1.897  1.429  1.076  0.763  0.431 
0.143 
0.606  0.945  1.023 
0.8 
4.460 
3.282  2.555  2.008  1.545  1.104  0.617 0.147  0.732  2.267 
0.9 
7.213 
5.688  4.446  3.717  2.964  2.322  1.651  0.884 0.154  1.452 
delta 
1.0 
12.216 
10.255  8.099  6.761  5.648  4.562  3.762  2.557  1.456 0.161 
Table6:Meanrelativeabsoluteerror(RAE)oftotalrevenueforecastsobtainedthrough2-parameter
forecastingmodelsanddifferentpairsof discountfactors.
Forecasting Accuracy
To test the forecasting accuracy of our models we followed a procedure similar to that used by
SawhneyandEliashberg(1996). Specifically,werandomlydivided ourdata setintoatrainingset
of50moviesandahold-outsetconsistingof the remaining30movies. Weusedthetrainingsetto
calibrateregressionequationsforP,Qandthenappliedtheequationstothehold-outsettoobtain
forecastsof amovie’stotalrevenueattheendofitsexhibitionhistory.
Table 6 lists the average relative absolute error (RAE = jPredicted¡Actualj=Actual) that
isassociated withthe above forecastsforeachof the 100 combinations of discountfactors –;" we
considered. Observe that forecasting errors are minimized when – = " and grow rapidly as the
twodiscountfactorsdiverge fromeachanother. Wewill,therefore,focusourattentiononthecase
–= ",corresponding tothe (highlighted)diagonal termsof Table 6. As the twodiscount factors
range between0 and 1,meanRAEfirstdeclines,reachesa minimum (14.1%)at– = "= 0:6 and
thenbeginstogrowagain. Interestingly,thecase–="=1,thatcorrespondstothestandardBass
model,hasameanRAEof16.1%. Thisis14%higherthanthemeanRAEof ourbestdiscounted
model. Our forecasting results, thus, confirm our theory-based hypothesis that introduction of
discountfactorstothetwotermsofa Bassequationimprovesthemodel’sforecasting accuracyin
thecontextofmovie revenues.
Of the two post-release motion picture forecasting models we are aware of, only the modelof
Sawhneyand Eliashberg(1996)isdirectlycomparable toours
15
. SawhneyandEliashberg (1996)
15
ThemodelofNeelameghamandChintagunta(1999)focusesonpredictingfirst-weekviewershipformoviesthat
22
SDK software service:VB.NET PDF Converter Library SDK to convert PDF to other file
This guide give a series of demo code directly for converting MicroSoft Office which users may quickly render and convert TIFF image file to PDF document.
www.rasteredge.com
SDK software service:C# PDF Converter Library SDK to convert PDF to other file formats
NET SDK library and components for converting PDF file in convert PDF to various document and image file formats Allow users to convert PDF to Text (TXT) file.
www.rasteredge.com
developed and tested BOXMOD-I, a model forforecasting the gross revenuesof motion pictures
based on their early box-office data. They tested how the forecasting accuracy of their model
improves as more box-office data becomes available and reported mean RAE of 71.1%, 51.6%,
13.2%,7.2%and1.8%when using no box-office data, one weekof data,two weeksof data, three
weeksof data and all available box-office data, respectively. Using only 3 daysof box-office and
userandcriticratingsdata,ourbestmodel(2-parametermodelwithdiscountfactors–="=0:6)
achieves levels of forecasting accuracy (mean RAE of 14.1%)forwhich BOXMOD-I requirestwo
weeksofboxofficedata. Thiscomparisonreinforcesouroriginalhypothesisthattheuseofonline
ratingsenablesreliable forecasts of the impactof anew experience goodto be made much faster
thanwitholdermethodologies
16
.
6 Summary, Managerial Implications, and Research Opportunities
Product review sites are widespread on the Internet and rapidly gaining popularity among con-
sumers. Previous research has established that online productratings have an influence on con-
sumer behavior (Chevalier and Mayzlin 2003; Senecal and Nantel 2004). This paper showsthat
thesesystemscanserveasavaluablesourceofinformationforfirmsaswell. Specifically,firmscan
use statisticsof online ratingsas areliable proxyof word-of-mouth inrevenueforecasting models
fornewexperiencegoods. Weapplythisideatothecontextofmotionpicturesandproposemotion
picture revenue-forecasting modelsthat use statisticsof online movie reviews posted byusers on
Yahoo! Moviesduring the first weekendof a new movie’s release to forecast thatmovie’s future
box-office performance.
Online movie ratings are available in large numbers within hoursof a new movie’s theatrical
release. Asapredictorofamovie’slong-termrevenueswehavefoundthemtobemoreinformative
thanothermeasurescurrentlyusedbyindustryexperts,suchascriticreviewsandearlycumulative
revenues. Theiruse,thus,allowsthegenerationof reliable forecastsmuchsooner. Specifically,we
haveshownthat,usingonlyopeningweekend(box-office,userandcriticreview)data,ourapproach
areintroducedsequentiallyindifferentmarkets(e.g. differentcountries).Theyusepost-releasedatafromonemarket
in order topredict the movie’s performance in another market. Their r objective, thus, is differentfromours: our
modelusesearlyboxofficeandratingsdatatopredictamovie’sfutureperformanceinthesame market.
16
SinceBOXMOD-Idoesnotincorporatecovariates,ourresultshould beinterpretedasevidenceforexplanatory
power of online ratingsrather than as astatement about the poweroftheunderlyingbehavioral modelon which
BOXMOD-Iisbased.
23
SDK software service:VB.NET Image: Tutorial for Converting Image and Document in VB.NET
VB.NET Project for Converting Image to Byte Array, Convert Word to Image in VB.NET Application. Use VB.NET Code to Convert Image to Stream, PDF to Image
www.rasteredge.com
SDK software service:C# PDF Convert to Word SDK: Convert PDF to Word library in C#.net
watermark and save PDF text, image, table, hyperlink and bookmark to Word without losing format. Powerful components for batch converting PDF documents in C#
www.rasteredge.com
cangenerateforecastswhoseaccuracywouldrequiretwoweeksof datausing oldertechniques.
Theabilitytoderiveearlypost-releaseforecastsofanewmovie’sperformancehastraditionally
beenof valuetoexhibitors(theaterowners). Exhibitorchainsneedtomanagetheyieldfromtheir
exhibitioncapacity,basedontheirestimatesofdemandformoviesthattheyarecurrentlyexhibiting.
Using such estimates theycan adaptthe exhibition capacityallocatedto a new movie, either by
droppingthe movie from atheaterorbyshiftingittoa smaller(orlarger)screeningroom. They
are, thus,veryinterested inearly forecastsofgrossbox-office revenues inmaking theirexhibition
decisions
17
. We argue that the ability to generate reliable forecasts so quickly after a movie’s
premiere can have important implications for motion picture marketing as well. Such knowledge
will allowmovie distributorstofine-tune amovie’scampaign or, perhaps, to developentirelynew
marketingstrategiesthatcanrespondto anaudience’sinitialreceptionof anewmovie
18
.
In addition to its managerial implications, our studyhas produced several empirical insights
relatedtotheuseofonlineproductratingsinrevenueforecasting.
First,wefoundthattheaveragevalenceofopeningweekenduserratingswasahighlysignificant
predictor of a movie’s long-term boxoffice performance. Given that the demographics of online
raters are skewed relative to the population of moviegoers, we also found that rebalancing the
averagevalenceofuserratings,bygivingequalweighttothearithmeticmeanofratingspostedby
malesandfemales,improvestheirpredictiveaccuracy.
Second,ouranalysisfounduserratingstobemoreinfluentialinpredictingfuturerevenuesthan
average professional critic reviews. Given the amount of attention that critic ratings have been
receiving until now, this result has considerable practical consequences. On the other hand, we
foundthecorrelationbetweentheuserandcriticratingstoberelativelylow;ourmodelswereable
toachievebetterforecastingaccuracybyconsideringaweightedaverageof userandcriticratings.
Thissuggeststhatadegreeofcomplementarityexistsbetweentheviewpointsofusersandexperts;
bothcan,thus,addvaluetopredictinganew product’sfuture success.
Third,wefoundthatthegenderdiversityofamovie’sonlineratersexhibitsapositivecorrelation
17
Todayexhibitorsusuallycommittoexhibitamovieforaminimumofthreetofourweeks. However,theincreasing
volatilityof second and later-weekrevenues (Lippman2003) plustheavailabilityofrapidforecastingtools,suchas
theonesweproposeinthispaper,mightleadtheindustrytoadoptmoreflexiblecontractsthatallowexhibitorsto
re-evaluatetheirdecisionsimmediatelyaftertheopeningweek.
18
SeeMahajan,MullerandKerin(1984)forsomeearlyideasonhowfirmscanadaptadvertisingpoliciestopositive
andnegativeword-of-mouth.
24
SDK software service:C# PDF Convert to HTML SDK: Convert PDF to html files in C#.net
source code for quick integration and converting PDF to HTML is a 100% clean .NET document image solution, which that are included in target PDF document file.
www.rasteredge.com
SDK software service:C# PDF Convert to Tiff SDK: Convert PDF to tiff images in C#.net
If using x86, the platform target should be x86. C# programming sample for PDF to Tiff image converting. // Load a PDF file. String
www.rasteredge.com
with thatmovie’slong-term revenues. Thisfinding supportsthe theorythatword-of-mouththat
ismore widelydispersedamongdifferentsocialgroupsismore effectiveandsuggeststheneed for
furtherresearchindevelopinggoodmeasuresof WOMdispersionfromonlinedata.
Fourth,wefoundthattheweeklyvolumeofonlineratingsexhibitshighcorrelationwithweekly
sales, suggestingthatpeople postratingssoonaftertheywatcha movie. Ourstudysupportsthe
hypothesis,commonlymadeindiffusiontheory,thattheimpactofword-of-mouthonfuturesalesis
proportionaltothevolumeofpastadoptersbutdoesnotfindanyspecialsignificanceofthevolume
ofonlineratingsbeyondthat.
We conclude bypointing outa numberof limitations of the current studyand associated op-
portunitiesforfuture research. First, incommonwith the majorityofpastworkinthisarea,our
models do not incorporate the impact of competition from othermovies. Such an enhancement
isnot possible withourcurrentdata set,since we don’thave completeboxofficeandproduction
dataforallmoviesplayingonallweeks. Second,ourobjectiveinthispaperwastogeneratefuture
revenue forecastsfrom asingle,earlymeasurementofboxoffice revenuesand online ratings. We,
thus,donothavetoworryaboutpotentialendogeneityissuesassociatedwiththeinterplaybetween
word-of-mouthandrevenues. Infuturework,we plantoexamineamodelthatusesmeasurements
ofrevenuesandratingsatmultiplepointsintimetoobtainmoreaccurateforecasts;insuchamodel,
endogeneitywillbe amore importantfactor,andwillbe dealtwith accordingly. Third, givenits
forecastingfocus,ourstudydidnotattempttoconsidertheimportantquestion ofwhetheronline
ratingsinfluence (as opposedto predict) future revenues. The perspective of the paper has been
thatonlineratingsofferfirmsavaluable,real-time“window”,thatallowsveryfastmeasurementof
whatconsumersthinkabout a new product, asopposed to a force that, in itself, influencescon-
sumerbehavior. Throughoutthepaperwehave,thus,beenverycarefulnottomakeanystatements
aboutcausality. Giventheincreasingpopularityofonline productreview sites,aninvestigationof
causalitywouldbeanexciting nextstepofthisline ofresearch.
References
Admati,A.,and P.Pfleiderer (2000)Noisytalk.com: Broadcasting Opinionsina NoisyEnvi-
ronment,WorkingPaper1670R,StanfordUniversity.
25
SDK software service:C# PDF Convert to Jpeg SDK: Convert PDF to JPEG images in C#.net
RasterEdge PDF to JPEG converting control SDK (XDoc.PDF for .NET) supports converting PDF document to JPEG image file in .NET developing platforms using simple
www.rasteredge.com
SDK software service:VB.NET PDF Convert to Word SDK: Convert PDF to Word library in vb.
VB.NET PDF Converting DLLs for PDF-to-Word. This is an example for converting PDF to Word (.docx) file in VB.NET program. ' Load a PDF document.
www.rasteredge.com
Ainslie,A.,Dreze,X.and Zufryden,F. (2003)ModelingMovieLifecyclesandMarketShare.
WorkingPaper.
Anderson, E.W. (1998)Customersatisfactionandwordofmouth. JournalofService1(1),5-17.
Austin,B.and Gordon,T. (1987)Movie Genres: Towarda Conceptualized Model and Stan-
dardizedDefinition. InB.Austin(ed.) CurrentResearchinFilm: Audience,Economics,and
Law,Vol. 3. AblexPublishingCo.,Norwood,NJ.
Banerjee,A. and Fudenberg,D. (2004) Word-of-mouth learning. Games and Economic Be-
havior 46(1)1-22.
Bass,F. (1969) A new product growth model for consumer durables. Management Science 15
(January): 215-227.
Basuroy,S.,Chatterjee,S.,and Ravid,S.A. (2003)How CriticalareCriticalReviews? The
BoxOfficeEffectsofFilmCritics,StarPowerandBudgets. JournalofMarketing67(October)
103-117.
Bowman,D.,and Narayandas, D. (2001) Managing customer-initiated contacts with manu-
facturers: The impact on share of category requirements and word-of-mouth behavior. J.
MarketingRes. 38,291-297.
Brown,J.J.,and Reingen,P. (1987) Social tiesandword-of-mouthreferral behavior. J. Con-
sumer Res. 14350-362.
Chatterjee,R. and Eliashberg,J. (1990)TheInnovationDiffusionProcessinaHeterogeneous
Population: AMicromodelingApproach. ManagementScience 36(9),1057-1079.
Chevalier, J.A.,and Mayzlin,D. (2003)TheEffectof WordofMouth onSales: OnlineBook
Reviews. YaleSOMWorkingPaperNo’s. ES-28 &MK-15.
DeSilva,I. (1998)ConsumerSelection of Motion Pictures. TheMotion PictureMega-Industry.
B.Litman. Boston,MA,Allyn&BaconPublishing,Inc.
DeVany,A.and Walls,W.D. (1999)Uncertaintyin the movie industry: Doesstarpowerre-
ducetheterroroftheboxoffice? J. CulturalEconom. 23(4): 285-318.
26
Dodds, J.andHolbrook,M. (1988)What’sanOscarWorth? AnEmpiricalEstimationofthe
Effectof Nominationsand Awards onMovie DistributionandRevenues. In B.Austin(ed.)
Current Research in Film: Audience, Economics, and Law, Vol. 4. AblexPublishing Co.,
Norwood,NJ.
Duan,W.,Gu,B.,andWhinston,A.B. (2005)Do Online ReviewsMatter? -An Empirical
InvestigationofPanelData. WorkingPaper,Universityof TexasatAustin.
Elberse,A.,and Anand,B.N. (2005)TheEffectivenessofPre-Release AdvertisingforMotion
Pictures. WorkingPaper,HarvardBusinessSchool.
Elberse,A. and Eliashberg,J. (2003)DemandandSupplyDynamicsforSequentuallyReleased
ProductsinInternationalMarkets: TheCase of MotionPictures. Marketing Science 22 (3)
329-354.
Eliashberg,J.and Sawhney,M.S. (1994)ModelingGoestoHollywood: PredictingIndividual
DifferencesinMovie Enjoyment. ManagementScience 40(9): 1151-1173.
Eliashberg,J.,andShugan,S.M. (1997) Film critics: Influencers or predictors? Journal of
Marketing 61(2): 68-78.
Eliashberg,J.,Jonker,J.,Sawhney,M.S.and Wierenga,B. (2000)MOVIEMOD:AnIm-
plementableDecisionSupportSystemforPre-ReleaseMarketEvaluationofMotionPictures.
MarketingScience 19(3)226-243.
Faber,R.and O’Guinn, T. (1984) Effect of Media Advertising and Other Sources on Movie
Selection. JournalismQuarterly 61: 371-377.
Godes,D.,and, Mayzlin D. (2004)UsingOnlineConversationstoStudyWordofMouthCom-
munication. MarketingScience23(4): 545-560.
Granovetter,M. (1973)Thestrengthof weakties. American JournalofSociology 78(6): 1360-
1380.
Jones,J.M.and Mason, C.H. (1990)The role of distribution in the diffusion of new durable
consumerproducts. Technicalworking paper,Marketing Science Institute, Cambridge, MA,
27
90-110.
Jones,J.M.and Ritz, C.J. (1991)Incorporatingdistributionintonewproductdiffusionmodels.
Internat. J. Res. Marketing 8(June): 91-112.
Krider,R.and Weinberg,C. (1998)CompetitiveDynamicsandtheIntroductionofNewProd-
ucts: TheMotionPictureTimingGame. JournalofMarketingResearch35(February): 1-15.
Lippman,J. (2003)Box-OfficeRecordsMaskSecond-Weekend Drops. TheWall StreetJournal,
June6,2003.
Litman,B.R.andAhn, H. (1998) Predicting Financial Success of Motion Pictures. In B.R.
Litman (ed.) TheMotion PictureMega-Industry. Allyn &Bacon Publishing Inc.,Needham
Heights,MA.
Litman,B.R.,and Kohl,A. (1989) Predicting financial sucess of motion pictures: The 80’s
experience. TheJournalofMediaEconomics 2(1): 35-50.
Litman,B.R. (1983)Predicting Successof TheatricalMovies: An EmpiricalStudy. Journalof
PopularCulture16(Spring): 159-175.
Liu,Y. (2004) Word-of-Mouth for Movies: Its Dynamics and Impact on Box Office Receipts.
WorkingPaper,December2004.
Mahajan, V.,Muller, E.,andBass, F.M. (1990). New ProductDiffusionModelsinMarket-
ing: AReviewandDirectionsforResearch. Journalof Marketing 54(January): 1-26.
Mahajan, V.,MullerE.,and Kerin,R.A. (1984)IntroductionStrategyforNewProductsWith
PositiveandNegativeWord-of-Mouth. ManagementScience30(December): 1389-1404.
Meade,N. (1984)TheUseofGrowthCurvesinForecastingMarketDevelopment-aReviewand
Appraisal. Journalof Forecasting 3,429-451.
Muñoz,L. (2003)High-TechWordofMouthMaimsMoviesinaFlash. LosAngelesTimes,August
17,2003.
Neelamegham,P., and Chintagunta,P. (1999) A Bayesian Model to Forecast New Product
PerformanceinDomesticandInternationalMarkets. MarketingScience 18(2): 115-136.
28
Ravid,S.A. (1999)Information,Blockbusters,andStars: AStudyof theFilmIndustry. Journal
ofBusiness 72(4): 463-492.
Reingen,P.,Foster,B.,Brown,J.J.,and Seidman,S. (1984)Brandcongruenceininterper-
sonalrelations: Asocialnetworkanalysis. J. Consumer Res. 11,1-26.
Reingen,P.,and Kernan,P. (1986) Analysis of referral networks inmarketing: Methods and
illustration. J.MarketingRes. 23370-378.
Reinstein,D.A.and Snyder,C.M. (2005)TheInfluence ofExpertReviewsonConsumerDe-
mandforExperienceGoods: ACaseStudyofMovieCritics. JournalofIndustrialEconomics,
forthcoming.
Richins,M.L. (1983)Negativeword-of-mouthbydissatisfied consumers: Apilotstudy. J.Mar-
keting 47,68-78.
Sawhey,M.S.and Eliashberg,J. (1996) A Parsimonious Model for Forecasting Gross Box-
OfficeRevenuesofMotionPictures. MarketingScience 15(2): 113-131.
Schindler, R.and Bickart,B. (2003)Published’WordofMouth’: Referable,Consumer-Generated
Informationonthe Internet, inC.Hauvgedt, K.Machleitand R. Yalch(eds.),OnlineCon-
sumer Psychology: Understandingand InfluencingBehavior in theVirtualWorld. Lawrence
ErlbaumAssociates.
Senecal,S.andNantel,J. (2004)The Influence of Online ProductRecommendationsonCon-
sumers’OnlineChoices. JournalofRetailing 80,159-69.
Sorensen,A.T. and Rasmussen,S.J. (2004)IsAnyPublicityGoodPublicity? ANoteonthe
ImpactofBookReviews. WorkingPaper,StanfordUniversity.
Zufryden,F.S. (1996) Linking advertising to box office performance of new film releases - A
marketingplanningmodel. Journalof AdvertisingResearch. 36(4): 29-41.
29
Appendix A: Comparison with Three-Parameter Models
Our2-parameterforecastingmodelisbasedontheassumptionthat,giveninfinitetimeintheaters,
the theoretical market potential of all movies is the same and equal to the entire population of
moviegoers. Theadvantageof ourassumptionisthatit avoidsthe hurdle of estimating the total
marketpotentialof each movie asa separate parameter. Amore conventional modelingapproach
would have beento assume thateach movie’smarketpotential M is equal to its total box office
revenuesandtofita3-parameter(P;Q;M)revenuemodelwiththisassumption. Forbenchmarking
purposes,thisappendixreportstheresultsof fittinga3-parametermodeltoourdata.
Asbefore,thenonlinearestimationstepwasverysuccessfulingeneratingcoefficientsP
i
;Q
i
;M
i
foreach movie. The nextstepconsistsindeveloping linearpredictionmodelsthatrelate eachset
of coefficientstoourcovariates. We follow the same variable selectionprocedure thatwe used to
generateour2-parameterpredictionmodels. Theresulting modelsaresummarizedinTable7.
Totalrevenues. Themostimportantchallengeinestimatinga3-parametermodelistheestima-
tionofaproduct’stotalmarketpotentialdirectlyfromcovariates. InspiredbytheworkofSorensen
andRasmussen(2004)onbookreviews,weexperimentedwiththe followingexponentialmodel:
M
i
=M
i0
exp(X
0
i
fl)"
i
(5)
whereM
i
denotesmoviei’stotalbox-office revenues,M
i0
denotesmoviei’sopening weekendrev-
enues,andX
i
isourvectorof covariates. Model(5)canbe estimatedbylinearregressionthrough
thefollowingequation:
LRAT =ln
µ
M
i
M
i0
=X
0
i
fl+u
i
(6)
Fitting equation (6) to our data resulted in a respectable adjusted R
2
of 0.84. Three covariates
weresignificant: Amongordinalcovariates,onlyaverageuserratings(BAVG)turnedouttobesig-
nificant,providingfurtherevidencefortheimportanceofearlyuserratingsinforecastingamovie’s
long-termrevenueprospects. Theothertwosignificantvariables(PG,SLEEPER)arecategorical.
Interestingly, all covariates thatwere significant in predicting LRAT were also significant in the
2-parameterregressionmodel forQ. This resultisintuitive, because bothQ andLRAT describe
30
Documents you may be interested
Documents you may be interested