Complexity measures of supervised classification problems

Ho, Tin Kam; Basu, Mitra

doi:10.1109/34.990132

Cited by 511 publications

(56 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, in order to describe the real complexity of the problem and since our eventual goal is to study the behavior of various data mining methods for prediction, we need to find other measures that are independent of such choices. Previous works highlight the idea that a single descriptor may not be sufficient (Tin Kam & Basu, 2002).…”

Section: Datasets Complexitymentioning

confidence: 99%

Predicting Win-Loss outcomes in MLB regular season games – A comparative study using data mining methods

Soto-Valero

2016

International Journal of Computer Science in Sport

View full text Add to dashboard Cite

Baseball is a statistically filled sport, and predicting the winner of a particular Major League Baseball (MLB) game is an interesting and challenging task. Up to now, there is no definitive formula for determining what factors will conduct a team to victory, but through the analysis of many years of historical records many trends could emerge. Recent studies concentrated on using and generating new statistics called sabermetrics in order to rank teams and players according to their perceived strengths and consequently applying these rankings to forecast specific games. In this paper, we employ sabermetrics statistics with the purpose of assessing the predictive capabilities of four data mining methods (classification and regression based) for predicting outcomes (win or loss) in MLB regular season games. Our model approach uses only past data when making a prediction, corresponding to ten years of publicly available data. We create a dataset with accumulative sabermetrics statistics for each MLB team during this period for which data contamination is not possible. The inherent difficulties of attempting this specific sports prediction are confirmed using two geometry or topology based measures of data complexity. Results reveal that the classification predictive scheme forecasts game outcomes better than regression scheme, and of the four data mining methods used, SVMs produce the best predictive results with a mean of nearly 60% prediction accuracy for each team. The evaluation of our model is performed using stratified 10-fold cross-validation.

show abstract

Section: Datasets Complexitymentioning

confidence: 99%

Predicting Win-Loss outcomes in MLB regular season games – A comparative study using data mining methods

Soto-Valero

2016

International Journal of Computer Science in Sport

View full text Add to dashboard Cite

show abstract

“…According to Ho and Basu [11], the difficulty of a classification problem can be attributed to three main aspects: the ambiguity among the classes, the complexity of the separation between the classes, and the data sparsity and dimensionality. Usually, there is a combination of these aspects.…”

Section: Complexity Indices For Describing Datamentioning

confidence: 99%

“…For such, we employ a series of statistic and geometric descriptors described in [11]. These indices account for the difficulty of a classification problem by analyzing some characteristics of the dataset and the predictive performance of some simple classification models induced using the dataset.…”

Section: Introductionmentioning

confidence: 99%

Effect of label noise in the complexity of classification problems

2015

View full text Add to dashboard Cite

“…On the other hand, assessing the behavior of the dataset ECU, establishing the most relevant attributes, and reducing any kind of data noise or inconsistency are vital actions [21]. For this, some filters are used, as explained in [22, 23, 24, and 25], allowing more trustworthiness in subsequent intrusion detection stages.…”

Section: ) Data Normalizationmentioning

confidence: 99%

Data Leakage Detection Using Dynamic Data Structure and Classification Techniques

Maldonado¹,

Byron²

2015

Inge CUC

View full text Add to dashboard Cite

Abstract-Data leakage is a permanent problem in public and private institutions around the world; particularly, identifying the information leakage efficiently. In order to solve this problem, this paper poses an adaptable data structure based on human behavior using all the activities executed within the computer system. When applying this structure, the normal behavior is modeled for each user, so in this way, detects any abnormal behavior in real time. Moreover, this structure enables the application of several classification techniques such as decision trees (C4.5), UCS, and Naive Bayes, these techniques have proven efficient outcomes in intrusion detection. In the testing of this model, a scenario demonstrating the proposal's effectiveness with real information from a government institution was designed so as to establish future lines of work.Keywords-Data Leakage, Data Structure, Decision Tree C4.5, UCS, Naive Bayes Resumen-La fuga de información es un problema que está presente en instituciones públicas y privadas alrededor del mundo. El principal problema que se presenta es identificar de forma eficiente el filtrado de la información. Para solucionar este problema en el presente trabajo desarrolla una estructura de datos adaptable al comportamiento humano, utilizando como base las actividades ejecutadas dentro del sistema informático. Al aplicar esta estructura se modela un comportamiento NORMAL de cada uno de los usuarios y de esta manera detecta cualquier comportamiento ANÓ-MALO en tiempo real. Además, permite la aplicación de varias técnicas de clasificación como los árboles de decisión (C4.5), UCS y Naive Bayes las cuales han demostrado un eficiente resultado en la detección de intrusiones. Para probar este modelo se ha diseñado un escenario que sirve para demostrar la validez de la propuesta con información real de una institución gubernamental y para acreditar líneas futuras de trabajo.Palabras Clave-Fuga de Información, Estructura de Datos,Árbol de decisión C4.5, UCS , Naive Bayes.

show abstract

Complexity measures of supervised classification problems

Cited by 511 publications

References 20 publications

Predicting Win-Loss outcomes in MLB regular season games – A comparative study using data mining methods

Predicting Win-Loss outcomes in MLB regular season games – A comparative study using data mining methods

Effect of label noise in the complexity of classification problems

Data Leakage Detection Using Dynamic Data Structure and Classification Techniques

Contact Info

Product

Resources

About