Training with Input Selection and Testing (TWIST) Algorithm: A Significant Advance in Pattern Recognition Performance of Machine Learning

Buscema, Massimo; Breda, Marco; Lodwick, Weldon A.

doi:10.4236/jilsa.2013.51004

Cited by 27 publications

(41 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This criterion increases the statistical probability that the two sub-samples are equally balanced during the genetic evolution because of the quasi-logarithmic increase of the optimization process. We have also demonstrated experimentally [6] that when there is no information in a dataset, the behaviors of the TWIST algorithm, the Training and Testing Random Splitting and the K-Fold Cross Validation are equivalent. Therefore, TWIST does not code noise to reach optimistic results [6].…”

Section: Twist Algorithmmentioning

confidence: 98%

“…We have also demonstrated experimentally [6] that when there is no information in a dataset, the behaviors of the TWIST algorithm, the Training and Testing Random Splitting and the K-Fold Cross Validation are equivalent. Therefore, TWIST does not code noise to reach optimistic results [6].…”

Section: Twist Algorithmmentioning

confidence: 98%

“…We have shown in many previous papers [6]- [19] that the TWIST algorithm outperforms the other splitting strategy (i.e. random distribution and K-Fold Cross Validation) in terms of results when they are applied to real medical data and also to classic datasets available from the UCI Machine Learning Repository [20].…”

Section: Twist Algorithmmentioning

confidence: 99%

See 2 more Smart Citations

Pattern Recognition for Flank Eruption Forecasting: An Application at Mount Etna Volcano (Sicily, Italy)

Brancato¹,

Buscema²,

Massini³

et al. 2016

OJG

View full text Add to dashboard Cite

The analysis involved successive steps. First, the TWIST algorithm was used to select the most predictive attributes associated with the flank eruption target. During his work, the algorithm TWIST selected 11 characteristics of the input vector: among them SO2 and CO 2 emissions, and also many other attributes whose linear correlation with the target was very low. A 5 × 2 Cross Validation protocol estimated the sensitivity and specificity of pattern recognition algorithms. Finally, different classification algorithms have been compared to understand if this pattern recognition task may have suitable results and which algorithm performs best. Best results (higher than 97% accuracy) have been obtained after performing advanced Artificial Neural Networks, with a sensitivity and specificity estimates over 97% and 98%, respectively. The present analysis highlights that a suitable monitoring dataset inferred hidden information about volcanic phenomena, whose highly non-linear processes are enhanced.

show abstract

Section: Twist Algorithmmentioning

confidence: 98%

Section: Twist Algorithmmentioning

confidence: 98%

See 1 more Smart Citation

Pattern Recognition for Flank Eruption Forecasting: An Application at Mount Etna Volcano (Sicily, Italy)

Brancato¹,

Buscema²,

Massini³

et al. 2016

OJG

View full text Add to dashboard Cite

show abstract

“…In Tomar and Agarwal [18] least square twin SVM (LSTSVM) based FS was proposed on the same dataset. Buscema et al [19] used training with input selection and testing (TWIST) algorithm for FS. Finally, Subbulakshmi et al [20] used extreme learning machine (ELM) to select the most relevant fewer features from the Stat log dataset.…”

Section: Fig 7 Accuracy Of Echocardiogram Dataset Using Different Cmentioning

confidence: 99%

Cuckoo inspired algorithms for feature selection in heart disease prediction

Usman

Yusof

Naim

2018

Int. J. Adv. Intell. Informatics

View full text Add to dashboard Cite

Heart disease is a predominant killer disease in various nations around the globe. However, this is because the default medical diagnostic techniques are not affordable by common people. This inspires many researchers to rescue the situation by using soft computing and machine learning approaches to bring a halt to the situation. These approaches use the medical data of the patients to predict the presence of the disease or not. Although, most of these data contains some redundant and irrelevant features that need to be discarded to enhance the prediction accuracy. As such, feature selection has become necessary to enhance prediction accuracy and reduce the number of features. In this study, two different but related cuckoo inspired algorithms, cuckoo search algorithm (CSA) and cuckoo optimization algorithm (COA), are proposed for feature selection on some heart disease datasets. Both the algorithms used the general filter method during subset generation. The obtained results showed that CSA performed better than COA both concerning fewer number of features as well as prediction accuracy on all the datasets. Finally, comparison with the state of the art approaches revealed that CSA also performed better on all the datasets.

show abstract

“…The split of original data set in two subsets has not been obtained with a simple random split but with an evolutionary algorithms called TWIST, recently described (35) and developed in a special noncommercial research software at the Semeion Research Center in Rome, Italy. TWIST algorithm.…”

Section: Statistical Analysesmentioning

confidence: 99%

Pregnancy risk factors in autism: a pilot study with artificial neural networks

Grossi¹,

Veggo

Narzisi

et al. 2015

Pediatr Res

View full text Add to dashboard Cite

Background: Autism is a multifactorial condition in which a single risk factor can unlikely provide comprehensive explanation for the disease origin. Moreover, due to the complexity of risk factors interplay, traditional statistics is often unable to explain the core of the problem due to the strong inherent nonlinearity of relationships. The aim of this study was to assess the frequency of 27 potential risk factors related to pregnancy and peri-postnatal period. Methods: The mothers of 45 autistic children and of 68 typical developing children completed a careful interview. Twentyfour siblings of 19 autistic children formed an internal control group. results: A higher prevalence of potential risk factors was observed in 22 and 15 factors in external control and internal control groups, respectively. For six of them, the difference in prevalence was statistically significant. Specialized artificial neural networks (ANNs) discriminated between autism and control subjects with 80.19% global accuracy when the dataset was preprocessed with TWIST (training with input selection and testing) system selecting 16 out of 27 variables. Logistic regression applied to 27 variables gave unsatisfactory results with global accuracy of 46%. conclusion: Pregnancy factors play an important role in autism development. ANNs are able to build up a predictive model, which could represent the basis for a diagnostic screening tool.a utism is a specific neurodevelopmental condition that typically displays qualitative socio-communicative impairment and restricted, stereotyped interests and activities (1). Although a large proportion of children with autism manifests abnormal development during the first year of life, 15-62% of them show a regression between 18 and 24 mo of age after a period of apparently typical development (2,3).Approximately 70% of individuals with autism present a variable degree of intellectual disability (4) and expressive/ receptive language can be absent or very insufficient (5). Other problems, not exclusive of autism, are attention-deficit and disturbed behaviors as etero-autolesivity. Thirty percent of children manifest epileptic seizures by late childhood or adolescence and 10% of cases are associated with several genetic disorders such as tuberous sclerosis, Angelman syndrome, phenylketonuria, and fragile X syndrome (6).The etiopathogenesis of autism is not yet understood; the prevalence is undoubtedly rising, and it is not clear if this increase is linked to the diagnostic improvement or to a greater susceptibility of the population to the disease. Many twins and family studies point out the importance of inherited predisposition to the disorder even though epidemiologic research suggest the strong contribution of prenatal and early postnatal environmental factors. In fact, genetic factors alone account for no more than 20-30% of all cases, whereas other 70-80% are the result of a complex interaction between environmental risk factors and inherited or de novo genetic susceptibility (7).Many pregnancies and o...

show abstract

Training with Input Selection and Testing (TWIST) Algorithm: A Significant Advance in Pattern Recognition Performance of Machine Learning

Cited by 27 publications

References 32 publications

Pattern Recognition for Flank Eruption Forecasting: An Application at Mount Etna Volcano (Sicily, Italy)

Pattern Recognition for Flank Eruption Forecasting: An Application at Mount Etna Volcano (Sicily, Italy)

Cuckoo inspired algorithms for feature selection in heart disease prediction

Pregnancy risk factors in autism: a pilot study with artificial neural networks

Contact Info

Product

Resources

About