A Predictive Model using Personality Traits: A Survey

A Predictive Model using Personality Traits: A Survey

1M. Tech Student, Dept of Computer Engineering and IT, VJTI College, Mumbai, Maharashtra, India

2Associate Professor, Dept of Computer Engineering and IT, VJTI College, Mumbai, Maharashtra, India ***

Abstract - Choosing the correct career is a crucial undertaking, but with the abundance of new career alternatives and opportunities that arise every day, it can be challenging.TheCSIRestimatesthatroughly40%ofstudents make the wrong profession choice as a result of their confusion. As a result, pupils' productivity declines as a result ofpoorselection.Inordertoavoidnegativeoutcomesbrought on by making the wrong option, it is crucial to select the right vocation at the right age. A crucial responsibility for the generation of today, as the world becomes increasingly technological, is choosing a career. This problem primarily affects pupils who are interested in different things. Everyone wants their children to become engineers or doctors, but nobody ever asks the child what he or she is interested in. Moreover, parents are often concerned about their children's futures. So, this software benefits both the parents and the students.

Key Words: Language Models, Personality Prediction, PsycholinguisticFeatures,NLP,MBTI.

1.INTRODUCTION

Withtheincreaseofstudyandinquiryinvariousfields,there are numerous new job opportunities in each subject. This further complicates the task of selecting a career for studentsconcentratingintenthortwelfthschool.Thecauses of this confusion may include self-incapacity and selfcharactertraitsignorance,ignoranceofthevariousoptions available,comparablepremiumsinmanyprofessions,lackof presentation, advertise bombardment, anticipated public activities, peer pressure, etc. Due to these issues, the understudy may decide on a job that is not a good fit for them, which could lead to work disappointment, subpar performance,stress,andothernegativeoutcomesincluding socialnegligence.So,thereneedstobeproperguidancefor the understudy's brain science, interest, and capacity to workinaparticulararea.

Oneofthekeyfactorsonwhichourfuturedependsisour career.Modernizationhasledtotheemergenceofseveral new job options, but on the flip side, an ever-growing number of graduates are graduating each year, boosting competitiveness. In order to provide the finest profession optionthatmatchestheirpersonalityandtalents,itisalso necessary to be aware of their hobbies, strengths, and weaknesses. Since we are in the information age, informationisconstantlygrowinginmanyareas,andmuch

beneficialknowledgeiscollectedfromthisdataandapplied tovariouschanges.

1.1 Personality Traits

Theterm"personalitytraits"isusedtodescribeasetof relativelyconsistentthought,feeling,andbehaviorpatterns thathavebeenlinkedtoavarietyofsignificantlifedecisions and consequences. Particularly, personality characteristics havefrequentlybeenlinkedtopersonaloutcomes(suchas happiness,psychopathology),interpersonaloutcomes(such asrelationshipsatisfaction),andsocialinstitutionaloutcomes (suchascareerchoices,worksuccess).Asaresult,thereis growinginterestincreatingmodelsthatcanautomatically predict people's levels of personality traits using internet data about human behavior and preferences (i.e., digital footprints).

NeedofPersonalityTraits?

1)Recommendersystems

2)Productandservicepersonalization

3)CareerAnalysis

4)Jobscreenings

5)Socialnetworkanalysis

6)Sentimentanalysis

1.2 Career Analysis

The choice of the best university or college no longer markstheconclusionofcareerplanning,whichwasoncea one-timeevent.Itcontinuesuntilwefindtheidealposition and a fulfilling work profile. One often chooses a stream based on their prior greatest achievement after finishing their education. Students that receive 90 percentiles in science desire to be engineersor doctors,andthesame is trueforcommerceandallothersubjects.Moreoftenthan not,itturnsouttobeahastyorpoordecision.Thecriteria bywhichweassessourcareerdecisionarestillunjustified.

Most Common Factors that Influence the Career Choice AmongStudents:

Parents’Desire/Parental

1)PressurePeerPersuasion

2)PastPerformance

3)

Potential

4)Placements

5)

PersonalityDriven

6)Affordability

1.3 MBTI

The Myers-Briggs Type Indicator (MBTI), used in personality typology, is an introspective self-report questionnaire that identifies various psychological preferences in how people view the world and make decisions. Although being extensively accepted by the scientificcommunity,itiswidelyviewedaspseudoscience. Introversionorextraversion,sensingorintuition,thinking or feeling, and evaluating or perceiving are the four categoriesthattheexamtriestorate.Tocreatefour-letter testresultslike"INTJ"or"ESFP,"oneletterisselectedfrom eachgroup.

calculate some data that is useful to predict the proper careers.Thissystemfunctionsasacareercounsellingsystem in which students submit their academic records, take an aptitudetest,andfilloutinformationabouttheirinterests, grades,andotherfactors.

VigneshSEtal.[2],inordertoforecastthebestdepartment for a person based on their skills as determined by an objective test, this article attempted to design a career system. One will automatically select the proper course if theycompletetheonlineevaluationwehaveprovidedinour system, which will help lower the failure rate due to selecting the incorrect career route. Dataset variety, appropriate career analysis, and improved visualization. Recentdatasetsandbetteralgorithmscouldbeemployed, however there are no personality traits. Techniques including HTML, CSS, Flask-API, Neural Networks, and Kmeans clustering were employed. For department recommendationpurposes, thesuccessrateineachofthe clusters is calculated and utilized to determine which clusterswillhaveahighersuccessrateandalowerfailure rate. In this project, a web-based application for career guidancehasbeenconceivedanddevelopedafterextensive researchonthetopic.

Fig-1: MBTItypes

2. LITERATURE SURVEY

AshutoshShankhdharEtal.[1],thisstudysoughttoenhance the recommendation process by helping readers choose a career path based on their personality traits, areas of interest, and readiness to enroll in a course. It also recommendedthebestuniversitiesforthemdependingon theirlocationandtuitioncosts.Academicperformancewas evaluated in order to determine whether the student possessedthenecessaryskillstofollowaspecificpredicted vocation. This also includes details about the student's gender,nationality,age,howmanydaystheyattendclass, whether they participate in extracurricular activities, whethertheypayattention inclass,andhowmanyonline courses they are capable of. In the disciplines of data, techniqueslikeKNN,NeuralNetworks,K-meansclustering, D-Tree,andmanyothercutting-edgealgorithmsareusedto

YashMehtaEtal.[3],Theauthorsofthisresearchsuggested a brand-new deep learning-based model for predicting language-based personality traits. This model includes employed language model embeddings and conventional psycholinguistic features as features. They have also examinedhowspecificpsycholinguisticcharacteristicsaffect a personality trait's ultimate prediction. According to the findings, psycholinguistic characteristics that model languagetypicallyoutperformthosethatdonot.Overall,the BERT-base+MLPmodeloutperformedBERT-large+MLPin terms of predicting the Big Five personality traits and the MBTIdimensions.Ourmodels'predictedperformanceson theEssaysdatasetandtheKaggledatasetoutperformedthe existingstate-of-the-artby1.6%and1%,respectively.Also, results from our interpretable machine learning study partiallyconcurwithearlierpsychologicalstudies.

KartikeyJoshiEtal.[4],Theyhavedevelopedasystemfor recommendingcareersthatwillassisttoday'syoungstersin decidingwhichprofessionalrouteisbestfortheirfutureand willyieldthebestoutcomesiftheytakethatpath.Thiswill enhance the student's performance and pique their attention,allowingthemtoconcentrateontheirdesiredjob. Thissystemisbuiltona testthatstudentsmusttake,and dependingontheirresponses,itwillproduceasummarized result. This system's primary goal is to give a general summaryoftheartificialintelligencemethodsweemployed to forecast the student's success. This system will also concentrateonhowweareidentifyingattributesinstudent databymeansofpredictionalgorithms.Usingthistechnique turned out to be advantageous for educators, educational

institutions,andstudentsalike.Thesystemwillproclaimthe finalconclusiondependingontheknowledgeweplacedinto itandtheartificialintelligencethatunderpinsit.Absenceof a personality test and accuracy dependent on particular featuresarebothflawsintheapproach.SVM,DecisionTree, Artificial Intelligence, and Student Career Prediction are someofthetermsthatwereemployedinthecreationofthis research.

Md. Yeasin Arafath Et al. [5], They used classification to examine survey data on successful alumni who are now employed,andtheyusedthatdatatoforecastthecareersof final-year students based on certain desirable characteristics. We primarily examined various academic, technical, and interpersonal aspects of the alumni's undergraduatecareersaswellastheircurrentlineofwork. Theircurrentlineofworkisregardedas aclasslabel,and thequalitytraitsareregardedasfeatures.Thisdataareused totrainthemodels,whichthenusetherepliesoftherunning studentswhohavefinishedtheirthirdyeartoforecasttheir careers based on the same quality criteria as test sets. Several classification methods are available. So, we used various classification algorithms and conducted a comparisonoftheclassifiers'performances.Theaccuracy, precision,recall,andf-measureofamodelareonlyafewof themetricsusedtoassessitsperformance.Severaldatasets areutilizedtoincreasediversity.Itmightbebetterifmore other ML methods were used. techniques for data mining likeID3,RandomForest,andCart.

ChanchalSumanEtal.[6],Theyhaveputupaparadigmfor resolving the issue of guessing a user's personality from videos. background, facial and from the user's video are takentheaudiofeatures.Themostoftenusedmeasurements in the literature are the big-five personality traits. Extraversion,Neuroticism,Agreeability,Conscientiousness, andOpennessarethesetraits.Manypracticalusesexistfor automatic personality characteristic prediction, including forensics,recommendersystems,personalizedservices,etc. Inthiswork,wehaveputforthaframeworkfortacklingthe issueofpredictingauser'spersonalitycharacteristicsfrom videos. From the user's video, ambient, face, and audio elementsareretrieved.ThedatasetmadepublicinChalearn16isusedtoassessthesystem'sperformance.Accordingto theresultsoftheexperiments,employingasmallnumberof the video's photos can yield better results than using the entirecollection.TheFac-modelvisual'sarchitectureisthe sameastheAmbvisual'smodelarchitectureforextracting features.

MadhuraJayaratneal.[7],InordertoestimateHEXACOtrait values from textual material, the authors created a regressionmodelusingnaturallanguageprocessing(NLP) and machine learning techniques. With an average correlationofr=0.39,theydiscoveredthattermfrequencyinversedocumentfrequency(TF-IDF)withLatentDirichlet

Allocation(LDA)themesperformedthebestwhencompared totheotherfivetextrepresentationtechniques.Incontrast, IBM'sPersonalityInsightsservice,developedusingtextdata fromTwitter,claimsanaveragecorrelationofr=0.31while big research using Facebook messages-based inference of Big5personalityshowedanaveragecorrelationofr=0.35. Theirresearchdemonstratesthatpersonalitytraitscanbe accurately inferred from the text of answers to common interview questions about past conduct and situational judgement.Inordertoself-rate people'spersonalities, we used information from over 46,000 job applicants who participatedinanonlinechatinterviewthatalsocontaineda personalityquestionnairebasedonthesix-factorHEXACO personalitymodel.

ZhanmingGuanEtal.[8],TheuniquemodelPersonality2vec, which is based on network representation learning, fully utilises the semantic, personality, and structural data in users'messages.Themodel trainsonthe network usinga novelbiasedwalkalgorithmandabetterskip-grammethod, anditeventuallyproducesapersonalityvectorforeachuser. Thepersonalityvectorsusedbytheauthoraretoforecast users' Big Five personality ratings using regression techniques.Usingthreepersonalitydatasets,experimental resultsdemonstratethatpersonality2vecbeatssevenwidely used approaches. First off, the majority of techniques primarily concentrate on the grammatical and semantic content of users' writing. Earlier techniques used psychologicaldictionarieslikeLIWCandMRC,whichwere createdbasedonpsychologicalandstatisticalexpertise,to extract linguistic properties. Some techniques attempt to employ NLP approaches to extract semantic features and combine those features with linguistic features for personalitypredictioninrecentyears,withtheemergenceof deeplearningandthematurationofNLPtechnology.

Z. Mushtaq al. [9], They suggested a system that could be erroneous because it asks consumers to fill out questionnairestoobtaintheirpersonalityinsights.andmake anefforttofilloutthesurveycarefully.Nevertheless,whenit comestosocialmedia,individualsdonotgivetheirideasany thoughtbeforepostingthem.Hence,theinformationgleaned from social media could be invaluable in identifying the personality type of the user. In this article, we suggest a methodforanalyzinguserdatafromsocialmediasites by fusingtwoalready-existingmachinelearningmethods,such as K-Means Clustering and Gradient Boosting, in order to identifyuserpersonalitytype.Also,thisstudycontributesto the analysis of the actual relationship between a user's personality and the data they post on social media. The Myers-Briggs Type Indicator (MBTI), developed by Swiss psychiatristCarlJung,wasusedinthisessay.Awell-known personality assessment tool used to determine a person's personalitytype,areasofstrength,andpersonalpreferences istheMyers-BriggsPersonalityIndicator.

Hans Christian Et al. [10], They have created a novel prediction that combines features from numerous pretrained language models, such as BERT, RoBERTa, and XLNet, with a multi-model deep learning architecture. Bidirectional context feature combined with extraction approach creates a prediction model for each attribute utilizing numerous social media data sources, including FacebookandTwitter.Findingshaveamaximumaccuracyof 86.2% and a 0.912 f1 measure score on the Facebook dataset,andanaccuracyof88.5%anda0.882f1measure score on the Twitter dataset. The tools used for model creationincludeBERT,RoBERTa,XLNet,NLPFeatures,and the Big Five personality model. A well-known personality model, known as the major fve personality traits, has frequentlybeen embracedintheliteratureasthedefacto standardforpersonalityevaluationduetoitssimplicityand shown effectiveness. Many techniques, some of which are basedonensembledmodelsanddeeplearning,canbeused toextractembeddedcontextualizedwordsfromtextualdata forpersonalitypredictionsystems.

3. PROPOSED SYSTEM

3.1 Problem Statement

To implement a Predictive Model for the Career Analysis usingMBTITraitandpsycholinguisticfeatures.

3.2 Problem Elaboration

The accuracy of prediction actually lies with the set of relevant skill parameters, interpersonal and academic factors.Thissystemworksasacareercounsellingsystemin which student give academic record and give the test of aptitudeandalsofilldetailssuchashobbies,marks,etcand thenthesystemwillrecommendasuitablecareerforthat student.Thissystemisalsohelpfultodeterminetheinternal traitsofthestudentwhichwillalsobeconsideredtofindan appropriatecareer.

3.3 Proposed Methodology

The system will ask the user or student to create a paragraph-styleessayinordertoanticipatetheirpersonality and then offer a job that fits their as well as taking into accountparticulardecisionsorpreferencesmade.Oneofthe mostrecent embeddingtechniquesoralgorithmsforlowdimensionalvectorspaceiscalledlongshorttermmemory (LSTM),whichmakestheproposedsystemperformbetter becauseitissimpleforthesystemtograsp.Theaccuracyof machinelearningmodelssuchasLongShort-TermMemory (LSTM),DecisionTree,KNN,andLogisticRegressionwillbe compared. The most accurate and efficient model among these will be taken into consideration. We will create correspondinganalysesforcorrespondingdomainsutilizing the 16 personalities from the MBTI Kaggle dataset. The modelwillnextbetrained(80%)andtested(20%)usingthe

aforementioned procedures. The dataset is then classified usingtheaforementionedtechniques.Thelogicformaking recommendationsisthendeveloped.Wewillchoosethebest modelafterevaluatingtheaccuracyofthethreemodels,and wewillthendevelopamodelforarecommendationsystem for each area. Hence, machine learning has assisted us in enhancingsystemperformanceanddevelopingasuperior recommendationsystem.

DataCollection:

DatafromtheKaggleMBTIdatasetisusedforallcareerand personalitycategories.Thisinformationwasgatheredfrom the PersonalityCafe forum and offers a wide variety of personsengagingincasualonlinesocialinteraction.Thelast 50 items a user wrote on the website are listed in 8675 entriestogetherwiththeirMBTIbinarypersonalitytypein thisdataset.Itisabout8.7GBinsize.BothJSONandpicture formats are supported. The JSON datasets will be transformed into CSV files. The appropriate dataset will containavarietyofdomain-relatedparameters.

TrainingandTestingthedataset:

Wewillfirstdopreprocessingonthedatasetstoeliminate duplicate, invalid, and null values from the dataset before startingthemodeltrainingprocedure.Therelevantdatasets will then be split into a training set and a testing set. Moreover,thetrainingtotestsetratiowillbe80:20.Three models will be used to classify the data. The predictive modelwillbeusedtobuildrecommendationpredictionlogic inoursuggestedsystem.Atestsetwillbeusedtotestthe model.Recall,precision,support,F1measure,accuracy,and other parameters will be used to evaluate the system's performance.

4. CONCLUSIONS

Whenastudenttakesatestandprovidesinformationabout theirgradesandinterests,theproposedcareersystemwill process the data and match them with the ideal career optionbasedontheirabilities.Italsotakesintoaccounttheir personalitytraits,aptitudetestresults,andfieldofinterest. This website served as a career counsellor, predicting careersbasedontestsandrecords.Notonlywouldthissave

timeandeffort,butitwillalsogiveamoreaccuratepicture ofone'spersonalityandjobprospects.

REFERENCES

[1] A.Shankhdhar,A.Agrawal,D.Sharma,S.Chaturvediand M. Pushkarna, "Intelligent Decision Support System UsingDecisionTreeMethodforStudentCareer,"2020 International Conference on Power Electronics & IoT Applications in Renewable Energy and its Control (PARC), Mathura, India, 2020, pp. 140-142, doi: 10.1109/PARC49193.2020.246974.M. Young, The TechnicalWriter’sHandbook.MillValley,CA:University Science,1989.

[2] S.Vignesh,C.ShivaniPriyanka,H.ShreeManjuand K. Mythili, "An Intelligent Career Guidance System using MachineLearning,"20217thInternationalConference onAdvancedComputingandCommunicationSystems (ICACCS), Coimbatore, India, 2021, pp.987-990, doi:10.1109/ICACCS51430.2021.9441978.

[3] Y.Mehta,S.Fatehi,A.Kazameini,C.Stachl,E.Cambria andS.Eetemadi,"Bottom-UpandTop-Down:Predicting PersonalitywithPsycholinguisticandLanguageModel Features,"2020IEEEInternationalConferenceonData Mining (ICDM), Sorrento, Italy, 2020, pp. 1184-1189, doi:10.1109/ICDM50108.2020.00146.

[4] K. Joshi, A. K. Goel and T. Kumar, "Online Career Counsellor System based on Artificial Intelligence: An approach,"20207thInternationalConferenceonSmart Structures and Systems (ICSSS), Chennai, India, 2020, pp.1-4,doi:10.1109/ICSSS49621.2020.9202024.

[5] M. Y. Arafath, M. Saifuzzaman, S. Ahmed and S. A. Hossain, "Predicting Career Using Data Mining," 2018 International Conference on Computing, Power and CommunicationTechnologies(GUCON),GreaterNoida, India, 2018, pp. 889-894, doi: 10.1109/GUCON.2018.8674995.

[6] ChanchalSuman,SriparnaSaha,AdityaGupta,Saurabh KumarPandey,PushpakBhattacharyya,Amulti-modal personality prediction system, Knowledge-Based Systems,Volume236,2022,107715,ISSN0950-7051, https://doi.org/10.1016/j.knosys.2021.107715.

[7] M.JayaratneandB.Jayatilleke,"PredictingPersonality UsingAnswerstoOpen-EndedInterviewQuestions,"in IEEE Access, vol. 8, pp. 115345-115355, 2020, doi: 10.1109/ACCESS.2020.3004002.

[8] Z. Guan, B. Wu, B. Wang and H. Liu, "Personality2vec: NetworkRepresentationLearningforPersonality,"2020

IEEEFifthInternationalConferenceonDataSciencein

Cyberspace(DSC),HongKong,China,2020,pp.30-37, doi:10.1109/DSC50466.2020.00013.

[9] Z.Mushtaq,S.AshrafandN.Sabahat,"PredictingMBTI PersonalitytypewithK-meansClusteringandGradient Boosting," 2020 IEEE 23rd International Multitopic Conference(INMIC),Bahawalpur,Pakistan,2020,pp.15,doi:10.1109/INMIC50486.2020.9318078.

[10] Christian, H., Suhartono, D., Chowanda, A. et al. Text basedpersonalitypredictionfrommultiplesocialmedia data sources using pre-trained language model and model averaging. J Big Data 8, 68 (2021). https://doi.org/10.1186/s40537-021-00459-1

[11] M. T. Apple, J. Falout and G. Hill, "The Relationship Between Future Career Self Images and English AchievementTestScoresofJapaneseSTEMStudents,"in IEEETransactionsonProfessionalCommunication,vol. 63, no. 4, pp. 372- 385, Dec. 2020, doi: 10.1109/TPC.2020.3029662.

[12] A. V. Kunte and S. Panicker, "Using textual data for PersonalityPrediction:AMachineLearningApproach," 2019 4th International Conference on Information Systems and Computer Networks (ISCON), Mathura, India, 2019, pp. 529-533, doi: 10.1109/ISCON47742.2019.9036220.

BIOGRAPHIES

RutujaTaraleiscurrentlypursuing M. Tech from VJTI COE, Mumbai. She has done her B.E. (Computer Engineering)fromXavierInstitute ofEngineering,Mumbai

Prof. Pramila M. Chawan, is workingasanAssociateProfessor in the Computer Engineering Department of VJTI, Mumbai. She has done her B.E.(Computer Engineering) and M.E.(Computer Engineering)fromVJTICollegeof Engineering, Mumbai University. She has 32 years of teaching experienceandhasguided80+M. Tech. projects and 100+ B. Tech. projects. She has published 140+ papers in the International Journals, 20+ papers in the National/International Conferences/Symposiums.Shehas worked as an Organizing Committee member for 21

International Conferences and 5 AICTE/MHRD sponsored Workshops/STTPs/FDPs.Shehas participated in 14 National/International Conferences. She has worked as NBACoordinatoroftheComputer Engineering Department of VJTI for 5 years. She had written a proposal under TEQIP-I in June 2004 for ‘Creating Central Computing Facility at VJTI’. Rs. EightCroreweresanctionedbythe WorldBankunderTEQIP-Ionthis proposal. Central Computing FacilitywassetupatVJTIthrough this fund which has played a key role in improving the teaching learningprocessatVJTI.