• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 229
  • 93
  • 20
  • 19
  • 15
  • 11
  • 7
  • 7
  • 5
  • 5
  • 4
  • 4
  • 3
  • 3
  • 2
  • Tagged with
  • 495
  • 417
  • 94
  • 83
  • 67
  • 58
  • 53
  • 50
  • 47
  • 43
  • 42
  • 41
  • 37
  • 36
  • 36
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
91

Development of novel unsupervised and supervised informatics methods for drug discovery applications

Mohiddin, Syed Basha, January 2006 (has links)
Thesis (Ph. D.)--Ohio State University, 2006. / Title from first page of PDF file. Includes bibliographical references (p. 172-185).
92

Generalization error rates for margin-based classifiers

Park, Changyi, January 2005 (has links)
Thesis (Ph. D.)--Ohio State University, 2005. / Title from first page of PDF file. Document formatted into pages; contains ix, 63 p.; also includes graphics (some col.). Includes bibliographical references (p. 60-63). Available online via OhioLINK's ETD Center
93

Methods for improving the reliability of semiconductor fault detection and diagnosis with principal component analysis

Cherry, Gregory Allan, January 1900 (has links) (PDF)
Thesis (Ph. D.)--University of Texas at Austin, 2006. / Vita. Includes bibliographical references.
94

Selected topics in statistical discriminant analysis

Ounpraseuth, Songthip T. Young, Dean M. January 2006 (has links)
Thesis (Ph.D.)--Baylor University, 2006. / Includes bibliographical references (p. 110-114).
95

Διαχωριστική ανάλυση - λογιστική παλινδρόμηση

Χουντής, Βασίλειος 07 July 2010 (has links)
Στην σημερινή εποχή είναι μεγάλη η ανάγκη να κατατάσσουμε παρατηρήσεις σε γνωστές ομάδες - πληθυσμούς καθώς επίσης και να κάνουμε προβλέψεις. Υπάρχουν πολλές μέθοδοι που κάνουν ή σκοπό έχουν να κατατάσσουν παρατηρήσεις. Στην διπλωματική εργασία περιγράφω δυο από τις σημαντικότερες μεθόδους που χρησιμοποιούνται ευρέως στην στατιστική, την διαχωριστική ανάλυση (discriminant analysis) και την λογιστική παλινδρόμηση (logistic regression). Στο πρώτο μέρος αναφέρω τι είναι η διαχωριστική ανάλυση, δίνω συνοπτικά μερικές εφαρμογές της μεθόδου και περιγράφω την διαφορά από την ανάλυση σε συστάδες. Στην συνέχεια αναλύω τον διαχωρισμό δυο πληθυσμών που ακολουθούν την κανονική κατανομή και τα κριτήρια που πρέπει να λάβουμε υπόψη. Στόχος μας είναι να κατασκευάσουμε μια συνάρτηση που θα διαχωρίζει όσο το δυνατόν καλύτερα τους δυο πληθυσμούς. Πρέπει να σημειώσουμε ότι δεν υπάρχει τέλειος διαχωρισμός, δηλαδή ενδέχεται η συνάρτηση να κατατάσσει λανθασμένα μια παρατήρηση σε μια από τις δυο ομάδες. Για αυτό πρέπει να λάβουμε υπόψη τα κόστη λανθασμένης κατάταξης και τις εκ των προτέρων πιθανότητες. Ο βέλτιστος διαχωρισμός θα πραγματοποιηθεί αν καταφέρουμε να ελαχιστοποιήσουμε το κόστος λανθασμένης κατάταξης. Στο τμήμα 3 βρίσκω την συνάρτηση κατάταξης όταν οι δυο πληθυσμοί έχουν ίσους πίνακες διασποράς (γραμμικός κανόνας κατάταξης) αλλά και όταν έχουν άνισες διασπορές (τετραγωνικός κανόνας κατάταξης). Εφόσον, έχω φτιάξει την συνάρτηση κατάταξης το επόμενο βήμα είναι να την αξιολογήσω. Περιγράφω δυο τρόπους αξιολόγησης (επικύρωσης), τον υπολογισμό του ρυθμού σφάλματος και την holdout διαδικασία. Στο τμήμα 5 αναφέρω την διαχωριστική ανάλυση του Fisher, τι υποθέσεις έκανε και πως κατάφερε να φτάσει στην ίδια συνάρτηση κατάταξης. Στην συνέχεια κάνω μια γενίκευση της διαχωριστικής ανάλυσης αν έχω g πληθυσμούς και δίνω το νέο τύπο της συνάρτησης κατάταξης όταν έχω ίσους και άνισους πίνακες διασποράς (γραμμικό – τετραγωνικό διαχωριστικό σκορ). Ερμηνεύω γεωμετρικά το γραμμικό διαχωριστικό σκορ. Στο τελευταίο τμήμα μελετάω την μέθοδο του Fisher όταν έχω g πληθυσμούς και αποδεικνύω μερικά θεωρήματα. Στο δεύτερος μέρος της διπλωματικής περιγράφω μια άλλη διαδικασία κατάταξης, την λογιστική παλινδρόμηση. Δίνω συνοπτικά μερικές εφαρμογές της μεθόδου και αναλύω πότε χρησιμοποιούμε αυτή την μέθοδο. Ξεκινώντας από το απλό γραμμικό μοντέλο παλινδρόμησης , αναφέρω τα προβλήματα που έχουμε τώρα που η μεταβλητή είναι δυαδική και πως τα αντιμετωπίζουμε, καταλήγοντας στην μορφή που έχει η απλή λογιστική συνάρτηση. Περιγράφω τις ιδιότητες της λογιστικής αποκρινόμενης συνάρτησης και πως προσαρμόζουμε το λογιστικό μοντέλο παλινδρόμησης χρησιμοποιώντας τους εκτιμητές μέγιστης πιθανοφάνειας. Κατόπιν δίνω την ερμηνεία του συντελεστή παλινδρόμησης και δίνω την μορφή της λογαριθμικής συνάρτησης πιθανοφάνειας όταν έχω επαναλαμβανόμενες παρατηρήσεις. Στο τμήμα 4 περιγράφω το πολλαπλό λογιστικό μοντέλο παλινδρόμησης και στο τμήμα 5 πως κατασκευάζεται το μοντέλο. Ελέγχω αν μπορούμε να παραλείψουμε μερικές προβλέπουσες μεταβλητές, χρησιμοποιώντας ένα στατιστικό που λέγεται μοντέλο απόκλισης, αλλά και από τον έλεγχο του λόγου πιθανοφάνειας. Προτού όμως χρησιμοποιήσω το μοντέλο στην πράξη εξετάζω την καταλληλότητα του, δηλαδή αν ικανοποιεί τις ιδιότητες της λογιστικής αποκρινόμενης συνάρτησης και αναζητώ τα outliers και τις παρατηρήσεις που έχουν την μεγαλύτερη επιρροή. Στα τμήματα 7 και 8 περιγράφω τα συμπεράσματα για τις παραμέτρους της λογιστικής παλινδρόμησης και για τον αποκρινόμενο μέσο, ενώ στο τμήμα 9 αναφέρω πως γίνεται η πρόβλεψη καινούριων παρατηρήσεων. Τελειώνοντας αναφέρω την πολύτομη λογιστική παλινδρόμηση και περιγράφω συνοπτικά τις ομοιότητες- διαφορές της διαχωριστικής ανάλυσης και της λογιστικής παλινδρόμησης. / -
96

Talento Motor : estudo dos indicadores somatomotores na seleção de escolares para o futebol

Santos, Fábio Rosa dos January 2013 (has links)
A identificação de talentos no esporte é um tema atual e amplamente discutido. Este debate ocorre no âmbito conceitual e metodológico. Especificamente para o futebol, esta é uma temática que tem gerado diversas discussões no meio acadêmico e na prática cotidiana dos clubes brasileiros. Assim, o objetivo deste estudo foi descrever o perfil de crianças e jovens praticantes de futebol e identificar indicadores de desempenho esportivo que permitam desenvolver parâmetros e metodologias para detecção de possíveis talentos motores para o futebol. As amostras foram compostas por 361 jovens de 10 a 13 anos de idade cronológica, constituídas por 188 escolares, aleatoriamente selecionadas do Banco de Dados do PROESPBr e por 173 atletas de futebol, selecionados por critério de acessibilidade, provenientes de dois clubes de futebol do Estado do Rio Grande do Sul (RS). Foi medida a estatura, envergadura, massa corporal, flexibilidade (sentar e alcançar com banco), força/resistência abdominal (abdominais por minuto), velocidade (20 metros), agilidade (quadrado de agilidade), força explosiva de membros inferiores (salto horizontal) e aptidão cardiorrespiratória (6 minutos). Para descrever o perfil antropométrico e de aptidão física foi utilizada a estatística descritiva (média e desvio padrão). Para a identificação das variáveis preditoras recorreu-se à Análise Discriminante e foi utilizado o pacote estatístico SPSS, versão 17.0. Os resultados demonstraram que os atletas apresentaram valores médios superiores aos escolares em todas as variáveis nas idades estudadas, com exceção da massa corporal, estatura e envergadura aos 11 anos de idade. Quanto as variáveis preditoras, aquelas que discriminaram o grupo de atletas do grupo de escolares foram: aos 10 anos de idade - flexibilidade, força/resistência abdominal, agilidade, velocidade e resistência aeróbia; 11 anos de idade - força resistência abdominal, agilidade, resistência aeróbia, velocidade; 12 anos de idade - resistência aeróbia, agilidade, velocidade, força resistência abdominal e força explosiva de membros inferiores e 13 anos de idade - resistência aeróbia, agilidade, flexibilidade, velocidade, força resistência abdominal. Os resultados demonstraram que a partir da utilização da bateria de testes do Projeto Esporte Brasil foi possível discriminar atletas de futebol de escolares e elaborar modelos capazes de identificar escolares com perfil de atleta. Assim, a utilização destes modelos podem compor os critérios de seleção de jovens para a modalidade do futebol e, ainda, atuar como instrumento auxiliar aos professores de Educação Física escolar. / Talent identification in sports is a current topic and it is widely discussed. This discussion occurs within conceptual and methodological. Specifically for soccer, this is a topic that has generated many discussions on academic level and in the daily practice of the Brazilian clubs. The aim of this study was to describe the profile of children and young people playing soccer and identify performance indicators sports which build parameters and methodologies for detecting potential talents engines for soccer. The sample was composed of 361 young people 10-13 years of chronological age, consisting of 188 students, randomly selected from the database of PROESP-Br and 173 soccer players selected by criteria of accessibility, from two soccer clubs of Rio Grande do Sul (RS) State. We measured height, arm span, body mass, flexibility (sit and reach with bench), abdominal strength / resistance (abdominal per minute), speed (20 meters), agility (agility square), explosive strength of lower limbs (horizontal jump) aerobic resistance (6 minutes). To describe anthropometric and physical fitness it was used descriptive statistics (mean and standard deviation). To identify the predictors we used the discriminant analysis. All analyses was done in statistical package SPSS, version 17.0. The results showed that athletes had higher values to schoolchildren in all variables and ages, with the exception of body mass, height and spread at 11 years old. The predictor variables, those discriminated group of athletes and students were: at 10 years of age - flexibility, abdominal strength / resistance, agility, speed and aerobic endurance, 11 years old - Strength abdominal endurance, agility, endurance aerobic speed; 12 years old - aerobic endurance, agility, speed, strength, abdominal strength and explosive strength of lower limbs and 13 years old - aerobic endurance, agility, flexibility, speed, strength, abdominal strength. The results showed that it was possible to discriminate between and students using the tests from Project Sport Brazil battery. Thus, we developed models capable of identifying students with athlete profile. The use of the models developed can compose the selection criteria of youth to the soccer sport and also act as an aid to teachers of Physical Education.
97

Talento Motor : estudo dos indicadores somatomotores na seleção de escolares para o futebol

Santos, Fábio Rosa dos January 2013 (has links)
A identificação de talentos no esporte é um tema atual e amplamente discutido. Este debate ocorre no âmbito conceitual e metodológico. Especificamente para o futebol, esta é uma temática que tem gerado diversas discussões no meio acadêmico e na prática cotidiana dos clubes brasileiros. Assim, o objetivo deste estudo foi descrever o perfil de crianças e jovens praticantes de futebol e identificar indicadores de desempenho esportivo que permitam desenvolver parâmetros e metodologias para detecção de possíveis talentos motores para o futebol. As amostras foram compostas por 361 jovens de 10 a 13 anos de idade cronológica, constituídas por 188 escolares, aleatoriamente selecionadas do Banco de Dados do PROESPBr e por 173 atletas de futebol, selecionados por critério de acessibilidade, provenientes de dois clubes de futebol do Estado do Rio Grande do Sul (RS). Foi medida a estatura, envergadura, massa corporal, flexibilidade (sentar e alcançar com banco), força/resistência abdominal (abdominais por minuto), velocidade (20 metros), agilidade (quadrado de agilidade), força explosiva de membros inferiores (salto horizontal) e aptidão cardiorrespiratória (6 minutos). Para descrever o perfil antropométrico e de aptidão física foi utilizada a estatística descritiva (média e desvio padrão). Para a identificação das variáveis preditoras recorreu-se à Análise Discriminante e foi utilizado o pacote estatístico SPSS, versão 17.0. Os resultados demonstraram que os atletas apresentaram valores médios superiores aos escolares em todas as variáveis nas idades estudadas, com exceção da massa corporal, estatura e envergadura aos 11 anos de idade. Quanto as variáveis preditoras, aquelas que discriminaram o grupo de atletas do grupo de escolares foram: aos 10 anos de idade - flexibilidade, força/resistência abdominal, agilidade, velocidade e resistência aeróbia; 11 anos de idade - força resistência abdominal, agilidade, resistência aeróbia, velocidade; 12 anos de idade - resistência aeróbia, agilidade, velocidade, força resistência abdominal e força explosiva de membros inferiores e 13 anos de idade - resistência aeróbia, agilidade, flexibilidade, velocidade, força resistência abdominal. Os resultados demonstraram que a partir da utilização da bateria de testes do Projeto Esporte Brasil foi possível discriminar atletas de futebol de escolares e elaborar modelos capazes de identificar escolares com perfil de atleta. Assim, a utilização destes modelos podem compor os critérios de seleção de jovens para a modalidade do futebol e, ainda, atuar como instrumento auxiliar aos professores de Educação Física escolar. / Talent identification in sports is a current topic and it is widely discussed. This discussion occurs within conceptual and methodological. Specifically for soccer, this is a topic that has generated many discussions on academic level and in the daily practice of the Brazilian clubs. The aim of this study was to describe the profile of children and young people playing soccer and identify performance indicators sports which build parameters and methodologies for detecting potential talents engines for soccer. The sample was composed of 361 young people 10-13 years of chronological age, consisting of 188 students, randomly selected from the database of PROESP-Br and 173 soccer players selected by criteria of accessibility, from two soccer clubs of Rio Grande do Sul (RS) State. We measured height, arm span, body mass, flexibility (sit and reach with bench), abdominal strength / resistance (abdominal per minute), speed (20 meters), agility (agility square), explosive strength of lower limbs (horizontal jump) aerobic resistance (6 minutes). To describe anthropometric and physical fitness it was used descriptive statistics (mean and standard deviation). To identify the predictors we used the discriminant analysis. All analyses was done in statistical package SPSS, version 17.0. The results showed that athletes had higher values to schoolchildren in all variables and ages, with the exception of body mass, height and spread at 11 years old. The predictor variables, those discriminated group of athletes and students were: at 10 years of age - flexibility, abdominal strength / resistance, agility, speed and aerobic endurance, 11 years old - Strength abdominal endurance, agility, endurance aerobic speed; 12 years old - aerobic endurance, agility, speed, strength, abdominal strength and explosive strength of lower limbs and 13 years old - aerobic endurance, agility, flexibility, speed, strength, abdominal strength. The results showed that it was possible to discriminate between and students using the tests from Project Sport Brazil battery. Thus, we developed models capable of identifying students with athlete profile. The use of the models developed can compose the selection criteria of youth to the soccer sport and also act as an aid to teachers of Physical Education.
98

Talento Motor : estudo dos indicadores somatomotores na seleção de escolares para o futebol

Santos, Fábio Rosa dos January 2013 (has links)
A identificação de talentos no esporte é um tema atual e amplamente discutido. Este debate ocorre no âmbito conceitual e metodológico. Especificamente para o futebol, esta é uma temática que tem gerado diversas discussões no meio acadêmico e na prática cotidiana dos clubes brasileiros. Assim, o objetivo deste estudo foi descrever o perfil de crianças e jovens praticantes de futebol e identificar indicadores de desempenho esportivo que permitam desenvolver parâmetros e metodologias para detecção de possíveis talentos motores para o futebol. As amostras foram compostas por 361 jovens de 10 a 13 anos de idade cronológica, constituídas por 188 escolares, aleatoriamente selecionadas do Banco de Dados do PROESPBr e por 173 atletas de futebol, selecionados por critério de acessibilidade, provenientes de dois clubes de futebol do Estado do Rio Grande do Sul (RS). Foi medida a estatura, envergadura, massa corporal, flexibilidade (sentar e alcançar com banco), força/resistência abdominal (abdominais por minuto), velocidade (20 metros), agilidade (quadrado de agilidade), força explosiva de membros inferiores (salto horizontal) e aptidão cardiorrespiratória (6 minutos). Para descrever o perfil antropométrico e de aptidão física foi utilizada a estatística descritiva (média e desvio padrão). Para a identificação das variáveis preditoras recorreu-se à Análise Discriminante e foi utilizado o pacote estatístico SPSS, versão 17.0. Os resultados demonstraram que os atletas apresentaram valores médios superiores aos escolares em todas as variáveis nas idades estudadas, com exceção da massa corporal, estatura e envergadura aos 11 anos de idade. Quanto as variáveis preditoras, aquelas que discriminaram o grupo de atletas do grupo de escolares foram: aos 10 anos de idade - flexibilidade, força/resistência abdominal, agilidade, velocidade e resistência aeróbia; 11 anos de idade - força resistência abdominal, agilidade, resistência aeróbia, velocidade; 12 anos de idade - resistência aeróbia, agilidade, velocidade, força resistência abdominal e força explosiva de membros inferiores e 13 anos de idade - resistência aeróbia, agilidade, flexibilidade, velocidade, força resistência abdominal. Os resultados demonstraram que a partir da utilização da bateria de testes do Projeto Esporte Brasil foi possível discriminar atletas de futebol de escolares e elaborar modelos capazes de identificar escolares com perfil de atleta. Assim, a utilização destes modelos podem compor os critérios de seleção de jovens para a modalidade do futebol e, ainda, atuar como instrumento auxiliar aos professores de Educação Física escolar. / Talent identification in sports is a current topic and it is widely discussed. This discussion occurs within conceptual and methodological. Specifically for soccer, this is a topic that has generated many discussions on academic level and in the daily practice of the Brazilian clubs. The aim of this study was to describe the profile of children and young people playing soccer and identify performance indicators sports which build parameters and methodologies for detecting potential talents engines for soccer. The sample was composed of 361 young people 10-13 years of chronological age, consisting of 188 students, randomly selected from the database of PROESP-Br and 173 soccer players selected by criteria of accessibility, from two soccer clubs of Rio Grande do Sul (RS) State. We measured height, arm span, body mass, flexibility (sit and reach with bench), abdominal strength / resistance (abdominal per minute), speed (20 meters), agility (agility square), explosive strength of lower limbs (horizontal jump) aerobic resistance (6 minutes). To describe anthropometric and physical fitness it was used descriptive statistics (mean and standard deviation). To identify the predictors we used the discriminant analysis. All analyses was done in statistical package SPSS, version 17.0. The results showed that athletes had higher values to schoolchildren in all variables and ages, with the exception of body mass, height and spread at 11 years old. The predictor variables, those discriminated group of athletes and students were: at 10 years of age - flexibility, abdominal strength / resistance, agility, speed and aerobic endurance, 11 years old - Strength abdominal endurance, agility, endurance aerobic speed; 12 years old - aerobic endurance, agility, speed, strength, abdominal strength and explosive strength of lower limbs and 13 years old - aerobic endurance, agility, flexibility, speed, strength, abdominal strength. The results showed that it was possible to discriminate between and students using the tests from Project Sport Brazil battery. Thus, we developed models capable of identifying students with athlete profile. The use of the models developed can compose the selection criteria of youth to the soccer sport and also act as an aid to teachers of Physical Education.
99

Geochemical and petrographic characterization of platreef pyroxenite Package p1, p2, p3 and p4 units at the akanani prospect area, bushveld Complex, South Africa

Mandende, Hakundwi January 2014 (has links)
>Magister Scientiae - MSc / This study is focused on the Akanani prospect area, approximately 25 km north-west of the town of Mokopane, Limpopo Province where exploration geologists at the study area have classified the ‘pyroxenitic’ units into P1, P2, P3 and P4 units upward in order of succession with height based on their textures, mineralogy and colour. The primary aim of this study is to distinguish the distinctive geochemical and mineralogical characteristics that can be used to identify each unit (P1 to P4) and in so doing create major geochemical, petrographic and mineralogical variables that will help or facilitate the exploration for and recovery of PGE and BMS mineralisation. Geochemical and mineralogical variation studies were carried out on the cores from ZF044, ZF045, ZF048, ZF057, ZF078, ZF082 and M0023, located in the Platreef at the Akanani Prospect area on the farms Moordkopje 813LR and Zwartfontein 814LR. Using a combination of various multivariate statistical techniques (factor, cluster and discriminant analysis) and mineralogical studies (CIPW norm, microprobe analysis, petrography), the outcomes of the study have demonstrated that the Platreef at Akanani comprise at least four lithological units i.e. the basal pyroxenite portion referred to as the P1 unit comprises chromitite, pyroxenites and feldspathic pyroxenites with associated Cr, TiO2, chromite, pyroxenes, hematite and Fe2O3, the mineralized section of the P2 unit is characterized by harzburgite, serpentinized harzburgite and in places orthopyroxenites are present consistent with high MgO and LOI contents, the feldspathic portion referred here as the P3 unit is characterized by a feldspathic pyroxenite containing higher Al2O3, Na2O, K2O, albite, hypersthene and SiO2 and the top most portion of the P4 unit comprising CaO, Diopside, ilmenite, anorthite, apatite and P2O5 that can be interpreted to have formed by three separate magma pulses. Considering the possibility that the P4 unit is a hybrid melt of assimilated Platreef that interacted with intruding Main Zone magma, this reduces the number of magma pulses to two. The classification of P1, P2, P3 and P4 units of the Platreef at Akanani shows that the criteria used by mining personnel to classify the four lithological units is not definitive and therefore are not highly reliable. Although various multivariate statistical techniques were employed relatively similar elemental associations were obtained highlighiting the importance of this approach. The strongly positive correlation between sulphides, PGEs and chromite at Akanani is consistent with an orthomagmatic deposit that had been disturbed by significant hydrothermal activity, while in places a good BMS-PGE relationship is commonly associated with the main chromitite stringers in P1. Mineral and whole rock compositions of silicate rocks highlight the strongly magnesian nature of the ultramafic P2 unit. Mineral chemistry studies of chromite, orthopyroxene, olivine, clinopyroxene and plagioclase are consistent with the multi- emplacement model. Convective exchange resulted in the enrichment of iron at the bottom of the stagnant chamber, while incompatible elements migrated upwards consistent with iron depletion with stratigraphic height. Injection of P1 magma and subsequent mixing with country rocks gave rise to the formation of chromitites and addition of plagioclase component to the intruding magma. A normal fractionation trend is suggested between P2 and P3 consistent with enrichment of MgO in P2 and enrichment of Al2O3, Na2O, SiO2 and K2O in P3. The An% of 84.4 of plagioclase coupled with CaO enrichment in P4 is suggestive of some Main Zone influence and can be interpreted as resulting from partial melting and recrystallization of P3 in response to the intrusion of the Main Zone magma is suggested for the formation of the P4 unit. There exists a good correlation between the modal mineralogy and mineral chemistry as determined optically, the norm as determined by the CIPW norm and the whole-rock geochemical results as determined by multivariate statistics and conventional methods.
100

Sex determination from the bones of the forearm in a modern South African sample

Barrier, Isabelle Linda Odile 05 August 2008 (has links)
With a large number of unidentified skeletal remains found in South Africa, the development of osteometric standards to determine sex from various bones is crucial. It is imperative that a forensic anthropologist have access to a variety of techniques both morphological and metric, which can be used to establish accurate demographic profiles from complete, fragmentary and/or commingled remains. Standards for South African populations are available for the cranium, humerus, pelvis, femur, tibia, calcaneus, and talus. No research has been done on bones of the forearm, even though they are known through international studies to exhibit sexual dimorphism. The purpose of this research was to develop discriminant function formulae to determine sex from the radius and ulna for a South African population. The sample consisted of 200 male and 200 female skeletons from the Pretoria Bone (University of Pretoria) and Raymond A. Dart (Witwatersrand University) collections. Sixteen standard anthropometric measurements were taken from the radius (9) and ulna (7) and subjected to stepwise and direct discriminant function analysis. Distal breadth, minimum midshaft diameter and maximum head diameter were the best discriminators of sex in the radius, while minimum midshaft diameter and olecranon breadth were selected for the ulna. Classification accuracy for the forearm ranged from 76% to 86%. In summary, the radius and ulna can be considered moderate discriminators for determining sex in a South African group. However, it is advised that whenever possible these formulae are used in conjunction with visual methods to determine sex from skeletal remains. Copyright 2007, University of Pretoria. All rights reserved. The copyright in this work vests in the University of Pretoria. No part of this work may be reproduced or transmitted in any form or by any means, without the prior written permission of the University of Pretoria. Please cite as follows: Barrier, ILO 2007, Sex determination from the bones of the forearm in a modern South African sample, MSc dissertation, University of Pretoria, Pretoria, viewed yymmdd < http://upetd.up.ac.za/thesis/available/etd-08052008-090115 / > E961/ag / Dissertation (MSc)--University of Pretoria, 2008. / Anatomy / unrestricted

Page generated in 0.1332 seconds