A data mining approach to process optimization without an explicit quality function

Chong, Il-Gyo; Albin, Susan L.

doi:10.1080/07408170601142668

Cited by 30 publications

(14 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Naturally, one of the most popular applications of PLS models is to transform original large-scale data into lower dimensional data to deal with highly correlated data between independent and dependent variables (Lakshminaraynan et al, 1997). In this process, several PLS factors are extracted to explain most of the variation in both independent and dependent variables (Chong et al, 2007). When a nonlinear relationship is implicitly assumed among dependent and independent variables, nonlinear PLS models can be estimated to construct nonlinear functional relationships using either neural networks or Gaussian kernel (Qin and The PLS method can be valuable as an alternative to well known data mining models to predict response variables or as a path model to understand structural relationships among records.…”

Section: Partial Least Square (Pls) Methodsmentioning

confidence: 99%

Churn management optimization with controllable marketing variables and associated management costs

Kim

Lee

Johnson

2013

Expert Systems with Applications

View full text Add to dashboard Cite

In this paper, we propose a churn management model based on a partial least square (PLS) optimization method that explicitly considers the management costs of controllable marketing variables for a successful churn management program. A PLS prediction model is first calibrated to estimate the churn probabilities of customers. Then this PLS prediction model is transformed into a control model after relative management costs of controllable marketing variables are estimated through a triangulation method. Finally, a PLS optimization model with marketing objectives and constraints are specified and solved via a sequential quadratic programming method. In our experiments, we observe that while the training and test data sets are dramatically different in terms of churner distributions (50% vs. 1.8%), four controllable variables in three marketing strategies significantly changed through optimization process while other variables only marginally changed. We also observe that the most significant variable in a PLS prediction model does not necessarily change most significantly in our PLS optimization model due to the highest management cost associated, implying differences between a prediction and an optimization model. Finally, two marketing models designed for targeting the subsets of custo mers based on churn probability or management costs are presented and discussed.

show abstract

Section: Partial Least Square (Pls) Methodsmentioning

confidence: 99%

Churn management optimization with controllable marketing variables and associated management costs

Kim

Lee

Johnson

2013

Expert Systems with Applications

View full text Add to dashboard Cite

show abstract

“…Na sequência, as observações são randomicamente divididas em duas porções: a porção de treino é utilizada para identificar as variáveis mais importantes e a porção de teste representa novas amostras a serem categorizadas. Recomenda-se uma proporção de 60% para porção de treino e 40% para porção de teste [21]. A regressão PLS é então aplicada na porção de treino, visando capturar a relação entre as variáveis independentes e dependente.…”

Section: Métodounclassified

Seleção De Variáveis Para Categorização De Amostras Químicas

Anzanello

2017

Eclet. Quim. J.

View full text Add to dashboard Cite

This paper presents a method to select the best variables to categorize chemical samples into two classes, say conforming or non-conforming. For that matter, PLS regression is combined with a data mining tool, the k-Nearest Neighbor classification technique, through an iterative variable selection process. The recommended subset of variables is chosen based on several criteria: sensitivity, specificity and percent of retained variables. When applied to two datasets related to wine analysis and one associated to QSAR, the proposed method significantly reduced the number of variables required for classification, while yielding superior categorization performance when compared to using all original variables.

show abstract

“…O banco de dados é randomicamente dividido em duas porções: treino (N tr ) e teste (N ts ), com N = N tr + N ts . Recomenda-se manter 60% das observações na porção de treino (CHONG; ALBIN; JUN, 2007).…”

Section: Métodounclassified

Seleção de variáveis para classificação de bateladas produtivas com base em múltiplos critérios

Anzanello

2013

Prod.

View full text Add to dashboard Cite

Processos industriais são frequentemente descritos por um elevado número de variáveis correlacionadas e ruidosas. Este artigo apresenta um método para seleção das variáveis mais relevantes para classificação de bateladas de produção valendo-se de múltiplos critérios de desempenho (sensibilidade e especificidade). As bateladas são categorizadas em duas classes (conforme ou não conforme, por exemplo). O método utiliza a regressão PLS (Partial Least Squares) para derivar um índice de importância das variáveis de processo. Um procedimento iterativo de classificação das bateladas e eliminação das variáveis é então conduzido. Por fim, uma medida de distância euclidiana ponderada é aplicada para selecionar o melhor subconjunto de variáveis. Ao ser aplicado em dados de processos industriais, o método proposto reteve, em média, 12% das variáveis originais, elevando a sensibilidade em 9%, de 0,78 para 0,85, e a especificidade em 20%, de 0,64 para 0,77. Estudos de simulação permitiram avaliar o desempenho do método frente a cenários distintos.

show abstract

A data mining approach to process optimization without an explicit quality function

Cited by 30 publications

References 15 publications

Churn management optimization with controllable marketing variables and associated management costs

Churn management optimization with controllable marketing variables and associated management costs

Seleção De Variáveis Para Categorização De Amostras Químicas

Seleção de variáveis para classificação de bateladas produtivas com base em múltiplos critérios

Contact Info

Product

Resources

About