Empirical Based Job Matching Model using Expert Human Knowledge: A Mixed-Methods Approach

The overarching goal of this research was to, using resumes and job descriptions, design a system that enables the classification of candidates and job positions in software development and data science-related areas as suitable or not suitable. This was achieved by identifying and modeling recruitment processes conducted by hiring personnel, for application in the automatic profiling of personnel.

Content

• 1_instruments_eval: this folder contains the JSON and LaTeX versions of the instruments for the evaluation experiment.
• 2_fuzzymeasuredata: this folder contains the measures that defined the regular fuzzy measure $\mu$ used in the Choquet integration. Also, contains the job seniority weights $W_j$, $W_m$, $W_s$.
• 3_db_eval_and_query: this folder contains the rankings and ranking cutoffs of the CVs made by the experiment's participants and the general metadata associated with each participant in a database format. Special Python functions are provided to query this database.
• 4_train_test_set: this folder contains the CV and job description dataset used during the training of the Siamese network. Each dataset includes its table schema.

1_instruments_eval description

The instruments are grouped using the set's names Greece, Trees, Constellations, Fruits. Each set contains six associated CVs. For instance, inside Greece folder there are the Delta, Kappa, Lambda, Sigma, Omega, and Tau CVs. The LaTeX version of the sets was made using a modified version Clean Resume Template by Vinayak Sharma on Overleaf. The ''Present'' in the CVs refers to November of 2023.

2_fuzzymeasuredata description

It contains two Excel files. The FSpace_PowerSet_Measure.xlsx file has the definition of the regular fuzzy measure $\mu$, and SenorityLevelsWeights.xlsx contains the job seniority weights junior ($W_j$), mid-level ($W_m$), and senior or higher ($W_s$).

3_db_eval_and_query description

Three files are included in the 3_db_eval_and_query folder. The first one is hlp_ranking_experiment.db, this was created in python using SQLite3, and contains the experiment information as can be seen in the ER diagram. The second file SQLite-ExampleQueries is a jupyter notebook with some examples of the database usage. The Third file contains a dictionary of all tables, explaning the meaning of each column.

Example cell:

# Assuming that a connection is established above
# The IDs you want to search for
decision_making_professional_id = 14
case_id = 'Greece'

# SQL query using placeholders for parameters
sql_query = """
SELECT er.*, cs.case_id, cs.subject_id
FROM exp_rank er
JOIN experiment e ON er.experiment_id = e.id
JOIN case_subject cs ON e.case_subject_id = cs.id
JOIN decision_making_professional dmp ON er.dmp_id = dmp.id
WHERE cs.case_id = ? AND dmp.id = ?;
"""
# Define your query with parameters
parameters = (case_id, decision_making_professional_id)
# Execute the query and fetch the result into a DataFrame
df = pd.read_sql_query(sql_query, conn, params=parameters)

Result:

id	experiment_id	dmp_id	rank	above_cutoff	case_id	subject_id
1	1	14	5	1	Greece	Delta
2	2	14	2	1	Greece	Kappa
3	3	14	3	1	Greece	Lambda
4	4	14	1	1	Greece	Sigma
5	5	14	4	1	Greece	Omega
6	6	14	6	1	Greece	Tau

Contact

For any queries, suggestions, or clarifications regarding these resources, you can contact the authors through the following means:

María Elena

• 📧 Email:
  • malena.mzn@outlook.com
  • a214200821@unison.mx
• 🔗 LinkedIn

Jordan Joel

• 🔗 LinkedIn

About the authors

• María Elena Martínez-Manzanares has a Bachelor's Degree in Mathematics obtained in 2018 from the Universidad de Sonora (Mexico), and a Bachelor's Degree in Mathematics Education obtained in 2023 from the Universidad Abierta y a Distancia de México. In 2019 she obtained an MSc in Mathematics from the Universidad de Sonora. Currently, she is finishing a Ph.D. at Universidad de Sonora, focusing her work on the stochastic control area. She is close to concluding her second MSc in Data Science at the same university with a focus on the NLP area. She has research experience on Markovian and semi-Markovian theoretical models and problems using mixed methods in different areas of applied mathematics and mathematics education. She worked as a data scientist in the telematics sector.
• Jordan Joel Urias Paramo has a Bachelor's Degree in Computer Science, obtained in 2019 from the Universidad de Sonora (Mexico). He currently works as a data engineer and is a Data Science Master's Degree student at the Universidad de Sonora.
• Julio Waissman-Vilanova has a Ph.D. in Automatic Systems from the Institut National Polytechnique de Toulouse (France). He is a full professor at Universidad de Sonora, Mexico, and his current research interests include Image Processing, Pattern Recognition and NLP.
• Gudelia Figueroa-Preciado has a Ph.D. in Mathematics from the Universidad de Sonora (Mexico). She is a full professor at Universidad de Sonora, Mexico, and her current research interests include Sampling Methods, Experimental Designs, Inferential Statistics.

Acknowledgments

We want to acknowledge the following people for their valuable contributions to this work:

• Ph.D. Cipriano Arturo Santos-Borbolla and MBA Francisco Enrique Andrade-López from the Operations Artificial Intelligence group (Oper.ai) at the Instituto Tecnológico y de Estudios Superiores de Monterrey (Mexico);
• Evelyn Mercedes Medina-García from the Master's Program in Social Sciences at the Universidad de Sonora (Mexico);
• Ph.D. Ramón Soto de la Cruz and Francisco Alejandro Bernal-Cañez from the Department of Mathematics at the Universidad de Sonora (Mexico);
• M.Sc. Irenisolina Antelo-López from the Doctoral Program in Sciences with a specialization in Educational Mathematics at the Universidad de Sonora (Mexico);
• Computer Scientist, Luis Fernando Sotomayor-Samaniego.

Thank you for your interest in this research.