Searcharxiv⌕ Search

arXiv subjects

Javier Trejos-Zelaya

Publications and source records attributed to Javier Trejos-Zelaya.

3 recordsLinked to original sources

NEVEN: An Open-Source Polyglot Infrastructure Bridging Statistical Computing and Spreadsheet Workflows for Reproducible Research

We present NEVEN, an open-source infrastructure that embeds R, Julia, Python, Quarto, and JavaScript visualization libraries into Excel through native worksheet formulas. Users run econometric estimators, solve differential equations, and compile reproducible reports using expressions such as =R.MR_Lineal(...), =J.Optimize(...), and =P.ai_call(...), without procedural programming or external application switching. The platform offers over 70 R statistical procedures, 70 high-level Julia procedures backed by more than 200 mathematical functions, and full Python 3 integration for statistical computing and numerical analysis. Each language runtime runs in an independent operating-system process connected to the Excel XLL host via Windows Named Pipes and Protocol Buffers. Fatal errors terminate only the affected child process, leaving Excel operational. A static security sandbox checks user code against 25-30 restricted patterns per language, blocking shell execution, file deletion, unauthorized network access, and dynamic evaluation through five bypass-prevention mechanisms validated by 133 tests. An embedded WebView2 engine displays interactive Plotly charts, D3.js network and hierarchical visualizations, Leaflet maps, pivot tables, LaTeX notation, PDF documents, and Pluto.jl reactive notebooks through a unified formula, =NEVEN.v(). The docked NEVEN Studio environment offers nine specialized modules, including a Chart Gallery with 325 templates, an in-memory DuckDB SQL console, a retrieval-augmented AI assistant, and a multi-language script editor with English and Spanish localization. Econometric estimates match published benchmarks to four decimal places, while Julia linear algebra decompositions achieve machine-precision accuracy (relative error below 10^-14). Released under GNU General Public License v3, NEVEN includes 342 automated unit, property-based, and integration tests.

stat.CO↗

Clustering Binary Data by Application of Combinatorial Optimization Heuristics

We study clustering methods for binary data, first defining aggregation criteria that measure the compactness of clusters. Five new and original methods are introduced, using neighborhoods and population behavior combinatorial optimization metaheuristics: first ones are simulated annealing, threshold accepting and tabu search, and the others are a genetic algorithm and ant colony optimization. The methods are implemented, performing the proper calibration of parameters in the case of heuristics, to ensure good results. From a set of 16 data tables generated by a quasi-Monte Carlo experiment, a comparison is performed for one of the aggregations using L1 dissimilarity, with hierarchical clustering, and a version of k-means: partitioning around medoids or PAM. Simulated annealing perform very well, especially compared to classical methods.

stat.ML↗

Clustering via Ant Colonies: Parameter Analysis and Improvement of the Algorithm

An ant colony optimization approach for partitioning a set of objects is proposed. In order to minimize the intra-variance, or within sum-of-squares, of the partitioned classes, we construct ant-like solutions by a constructive approach that selects objects to be put in a class with a probability that depends on the distance between the object and the centroid of the class (visibility) and the pheromone trail; the latter depends on the class memberships that have been defined along the iterations. The procedure is improved with the application of K-means algorithm in some iterations of the ant colony method. We performed a simulation study in order to evaluate the method with a Monte Carlo experiment that controls some sensitive parameters of the clustering problem. After some tuning of the parameters, the method has also been applied to some benchmark real-data sets. Encouraging results were obtained in nearly all cases.

stat.ML↗