| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 3.42 MB | Adobe PDF |
Authors
Abstract(s)
O estudo teve por objetivo construir um modelo para identificar os benefícios e as funcionalidades
na utilização da Inteligência Artificial (IA), do tipo Rede Neuronal Artificial (RNA) múltiplas camadas
feedforward, aplicada no processo de recrutamento e seleção (R&S), de profissionais do cargo de
técnico de vendas, como um sistema de apoio à tomada de decisão, para auxiliar o profissional de
Recursos Humanos (RH). Esta pesquisa quantitativa exploratória utilizou técnicas estatísticas
descritiva, inferencial e o modelo de RNA. Utilizou-se dados secundários (14 atributos técnicos)
contidos em 800 currículos (CV) de indivíduos que se candidataram ao cargo, no período de 2014
a 2018. O algoritmo de treino priorizado foi o Bayesian Regularization; com funções de ativação
Tangente Simoidal e Linear e arquitetura com dois nós na camada escondida. Dos 800 indivíduos,
54,4% são do sexo masculino; 29,3% possui idade entre 24 e 29 anos; 63,1% possui 2.º grau
completo; 16,4% possui até 2 anos de experiência em vendas e 70,5% não possui formação
complementar em vendas. Verificou-se que há diferença nas médias das notas atribuídas ao CV por
sexo e que há associação entre a nota atribuída ao CV e o tempo de experiência. O modelo
apresentou excelentes resultados, na maioria dos casos a nota de classificação do currículo
atribuída foi igual ao alvo da classificação e o erro médio absoluto foi de 0,292 pontos, numa escala
de 1 a 10 pontos. A RNA aplicada ao R&S do referido cargo reduz o tempo gasto na atividade,
facilita o trabalho do profissional de RH e possibilita que este dedique mais tempo às etapas de
análise das competências comportamentais, como as dinâmicas de grupo e entrevistas. Além disto,
facilita a comunicação entre candidato e empresa, proporcionando que o feedback seja dado de
modo mais ágil.
The study aimed to build a model to identify the benefits and functionalities in the use of Artificial Intelligence (AI), of the type Artificial Neuronal Network (ANN) multiple layers feedforward, applied in the process of recruitment and selection (R&S), of professionals of the position of sales technician, as a support system for decision making, to assist the Human Resources (HR) professional. This quantitative exploratory research used descriptive, inferential statistical techniques and the RNA model. Secondary data (14 technical attributes) were used in 800 CVs of individuals who applied for the position, from 2014 to 2018. The priority training algorithm was Bayesian Regularization; with activation functions Tangent Simoidal and Linear and architecture with two nodes in the hidden layer. Of the 800 individuals, 54.4% are male; 29.3% are aged between 24 and 29 years; 63.1% have 2nd. complete degree; 16.4% have up to 2 years of sales experience and 70.5% have no further training in sales. It was found that there is a difference in the average grades attributed to the CV by sex and that there is an association between the grade attributed to the CV and the length of experience. The model presented excellent results, in most cases the rating score of the curriculum attributed was equal to the target of the classification and the average absolute error was 0.292 points, on a scale of 1 to 10. The RNA applied to the R&S of that position reduces the time spent on the activity, facilitates the work of the HR professional and allows him to dedicate more time to the stages of analysis of behavioural skills, such as group dynamics and interviews. In addition, it facilitates communication between candidate and company, providing that feedback is given in a more agile way.
The study aimed to build a model to identify the benefits and functionalities in the use of Artificial Intelligence (AI), of the type Artificial Neuronal Network (ANN) multiple layers feedforward, applied in the process of recruitment and selection (R&S), of professionals of the position of sales technician, as a support system for decision making, to assist the Human Resources (HR) professional. This quantitative exploratory research used descriptive, inferential statistical techniques and the RNA model. Secondary data (14 technical attributes) were used in 800 CVs of individuals who applied for the position, from 2014 to 2018. The priority training algorithm was Bayesian Regularization; with activation functions Tangent Simoidal and Linear and architecture with two nodes in the hidden layer. Of the 800 individuals, 54.4% are male; 29.3% are aged between 24 and 29 years; 63.1% have 2nd. complete degree; 16.4% have up to 2 years of sales experience and 70.5% have no further training in sales. It was found that there is a difference in the average grades attributed to the CV by sex and that there is an association between the grade attributed to the CV and the length of experience. The model presented excellent results, in most cases the rating score of the curriculum attributed was equal to the target of the classification and the average absolute error was 0.292 points, on a scale of 1 to 10. The RNA applied to the R&S of that position reduces the time spent on the activity, facilitates the work of the HR professional and allows him to dedicate more time to the stages of analysis of behavioural skills, such as group dynamics and interviews. In addition, it facilitates communication between candidate and company, providing that feedback is given in a more agile way.
Description
Mestrado APNOR e Mestrado de dupla diplomação com a UNIFACS - Universidade Salvador
Keywords
Inteligência artificial Recrutamento e seleção Recursos humanos Rede neuronal artificial