On the statistical evaluation of algorithmic's computational experimentation with infeasible solutions

Carvalho, Iago A.

doi:10.1016/j.ipl.2018.11.003

Cited by 6 publications

(6 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A series of statistical analyses were performed to investigate if there exists an algorithm that outperforms the others in predicting the consumer responses. To do so, we ranked the data presented in Table 3 using the ranking methodology proposed by Carvalho (2019). For every fruit and each consumer response, the results of the algorithms were ranked lexicographically, whereas the R 2 value had greater importance than the RMSE value.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

A comparison of machine learning algorithms for predicting consumer responses based on physical, chemical, and physical–chemical data of fruits

Ribeiro

Carvalho

Ferreira

et al. 2022

Journal of Sensory Studies

Self Cite

View full text Add to dashboard Cite

Machine learning algorithms are widely used for predicting the consumer response to several food products. Recent studies in the literature demonstrated that it is possible to predict the consumer response to fruits using the physical, chemical, and physical–chemical data of fruits as input for the machine learning algorithms. However, a myriad of machine learning algorithms exists, and there is no consensus on which algorithm is the best for this task. This work evaluates and compares the results of six of the most used machine learning algorithms, the Random Forest, the Decision Tree, the Support Vector Machine, the Multilayer Perceptron neural network, the K‐Nearest Neighbors, and the Multivariate Linear Regression, in predicting the consumers’ acceptance, expectation, and their ideal of sweetness, succulency, and acidity for three different fruits. The results demonstrated that, indeed, there is no algorithm that outperforms all others for this task. Every algorithm has its advantages and disadvantages and performs differently according to the fruit and the corresponding dataset. Therefore, it highlights the importance of carefully selecting, optimizing, and comparing several algorithms when one is interested in predicting the consumer response to fruits. Practical applications Fruits are mostly commercialized without a strong assurance of their quality, which is their physical–chemical and sensory aspects. The loss of control of these aspects may impact the consumer, which can acquire low‐quality products and, thus, lead to unsatisfaction. This research shows that machine learning algorithms can be employed to effectively predict the consumer's sensory response to fruits. The use of these algorithms, which are based on easy‐to‐obtain physical and physical–chemical data, can improve the quality control of fruits in the market. Therefore, fruit producers and markets can commercialize their products based on their quality, thus providing a better experience to the consumer, which, in turn, can improve their satisfaction.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Regression Modelsmentioning

confidence: 99%

A comparison of machine learning algorithms for predicting consumer responses based on physical, chemical, and physical–chemical data of fruits

Ribeiro

Carvalho

Ferreira

et al. 2022

Journal of Sensory Studies

Self Cite

View full text Add to dashboard Cite

show abstract

“…SBA was first proposed for the Minmax regret Weighted Set Covering problem under Interval Uncertainties [18,19] and later extended to M-BIP [38] and other minmax regret optimization problems under interval uncertainties [10,39,40]. It consists in investigating a set Q = {q 1 , q 2 , .…”

Section: The Scenario-based Algorithmmentioning

confidence: 99%

Fix-and-optimize metaheuristics for minmax regret binary integer programming problems under interval uncertainty

2022

View full text Add to dashboard Cite

The Binary Integer Programming problem (BIP) is a mathematical optimization problem, with linear objective function and constraints, on which the domain of all variables is {0,1}. In BIP, every variable is associated with a determined cost coefficient. The Minmax regret Binary Integer Programming problem under interval uncertainty (M-BIP) is a generalization of BIP in which the cost coefficient associated to the variables is not known in advance, but are assumed to be bounded by an interval. The objective of M-BIP is to find a solution that possesses the minimum maximum regret among all possible solutions for the problem. In this paper, we show that the decision version of M-BIP is Σ p 2 -complete. Furthermore, we tackle M-BIP by both exact and heuristic algorithms. We extend three exact algorithms from the literature to M-BIP and propose two fix-and-optimize heuristic algorithms. Computational experiments, performed on the Minmax regret Weighted Set Covering problem under Interval Uncertainties (MWSCP) as a test case, indicates that one of the exact algorithms outperforms the others. Furthermore, it shows that the proposed fix-andoptimize heuristics, that can be easily employed to solve any minmax regret optimization problem under interval uncertainty, are competitive with ad-hoc algorithms for the M-WSCP.

show abstract

“…In the second step, Friedman's test (Friedman, 1937) is used to verify whether there is a statistical significant difference between at least two of the evaluated heuristics. The ranking used by Friedman's test is computed using the biobjective lexicographical ranking scheme of Carvalho (2019). The null hypothesis of Friedman's test is that AMU, SBA, VND, and FO-HD have the same average relative optimality gap.…”

Section: Comparison With the Best Heuristics In The Literaturementioning

confidence: 99%

A fix‐and‐optimize heuristic for the minmax regret shortest path arborescence problem under interval uncertainty

Carvalho

Noronha

Duhamel

et al. 2021

Int Trans Operational Res

Self Cite

View full text Add to dashboard Cite

The minmax regret shortest path arborescence (M-SPA) problem under interval uncertainty is a robust counterpart of the shortest path (SP) arborescence problem, where arc costs are modeled as intervals of possible values. This problem finds applications in the design of topologies for low-power wireless personal area networks. The previous work presented a mixed-integer linear programming formulation and heuristics for this problem. In this paper, we propose a new heuristic for M-SPA, called the fix-and-optimize through heuristic decomposition (FO-HD). It consists of two steps. The first step solves multiple instances of the minmax regret SP problem under interval uncertainty, while the second step combines their solutions into an arborescence. As far as we know, this is the first work in the literature to apply the fix-and-optimize metaheuristic to minmax regret problems. We also show that a lower bound for the optimal solution of M-SPA can be computed from the first step of FO-HD. This lower bound allows us to show that the maximum average optimality gap of this heuristic was 0.2% among all classical and novel instances used in the computational experiments.

show abstract

On the statistical evaluation of algorithmic's computational experimentation with infeasible solutions

Cited by 6 publications

References 14 publications

A comparison of machine learning algorithms for predicting consumer responses based on physical, chemical, and physical–chemical data of fruits

A comparison of machine learning algorithms for predicting consumer responses based on physical, chemical, and physical–chemical data of fruits

Fix-and-optimize metaheuristics for minmax regret binary integer programming problems under interval uncertainty

A fix‐and‐optimize heuristic for the minmax regret shortest path arborescence problem under interval uncertainty

Contact Info

Product

Resources

About