Techniques of Decision Tree Induction

Grąbczewski, Krzysztof

doi:10.1007/978-3-319-00960-5_2

Cited by 9 publications

(9 citation statements)

References 127 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As a result, the classification of the instances continues down to each internal node based on the value of the characteristic determined in the test. If a leaf, a node with a single incoming edge, is reached then the classification is decided by its class [ 36 ]. The main advantage of DT is that it can be visualized easily and is relatively simple to interpret.…”

Section: Methodsmentioning

confidence: 99%

“…The DT algorithm can classify both numerical and categorical data while other algorithms can only handle one type of variable. DT has promising results in terms of minimum computation time and thus considered reliable for real-time systems [ 36 ]. The C4.5 decision tree was proposed by Ross Quinlan [ 37 ] and its rule generation has significantly sped up the training procedure.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Multirate Processing with Selective Subbands and Machine Learning for Efficient Arrhythmia Classification

Qaisar

Mihoub

Krichen

et al. 2021

Sensors

View full text Add to dashboard Cite

The usage of wearable gadgets is growing in the cloud-based health monitoring systems. The signal compression, computational and power efficiencies play an imperative part in this scenario. In this context, we propose an efficient method for the diagnosis of cardiovascular diseases based on electrocardiogram (ECG) signals. The method combines multirate processing, wavelet decomposition and frequency content-based subband coefficient selection and machine learning techniques. Multirate processing and features selection is used to reduce the amount of information processed thus reducing the computational complexity of the proposed system relative to the equivalent fixed-rate solutions. Frequency content-dependent subband coefficient selection enhances the compression gain and reduces the transmission activity and computational cost of the post cloud-based classification. We have used MIT-BIH dataset for our experiments. To avoid overfitting and biasness, the performance of considered classifiers is studied by using five-fold cross validation (5CV) and a novel proposed partial blind protocol. The designed method achieves more than 12-fold computational gain while assuring an appropriate signal reconstruction. The compression gain is 13 times compared to fixed-rate counterparts and the highest classification accuracies are 97.06% and 92.08% for the 5CV and partial blind cases, respectively. Results suggest the feasibility of detecting cardiac arrhythmias using the proposed approach.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Multirate Processing with Selective Subbands and Machine Learning for Efficient Arrhythmia Classification

Qaisar

Mihoub

Krichen

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…The unsupervised greedy stepwise forward attribute evaluation was performed with 1,449 instances to determine the most important parameters in predicting P. xylostella abundance. The M5P pruned decision-tree algorithm output is a flowchart with a tree structure, where each internal node denotes a test on an attribute, each branch representing a result of the test, and each leaf node holding a class label (Barros, de Carvalho, & Freitas, 2015;Grabczewski, 2014;Sharma, Kumar, & Maheshwari, 2015). M5P was selected since it is ideal for use with numeric data and was superior to the general linear model when the two methods were compared (Sharma et al, 2015).…”

Section: M5p Pruned Decision-tree Algorithm Induction and Hyperparam-mentioning

confidence: 99%

Disentangling factors limiting diamondback moth, Plutella xylostella (L.), spatio‐temporal population abundance: A tool for pest forecasting

Machekano

Mutamiswa

Mvumi

et al. 2019

J Applied Entomology

View full text Add to dashboard Cite

Data-mining techniques play an important role in hyperparameter optimization of heterogeneous environmental factors and their relative contribution as determinants of incidences in insect pest ecological studies. A multidimensional field-based surveillance was conducted in two seasons (24 months), July-June of each season (2015/2016 -season 1 and 2016/2017 -season 2) using sex-pheromone-baited traps and Thermocron i-Buttons to identify key determinants of population abundance of diamondback moth, Plutella xylostella L., across spatial horticultural hotspots of Botswana. The moth is a notorious global brassica pest. Pearson's product moment correlation matrix showed month of the year (M), mean temperature (T mean ) and maximum temperature (T max ) as positively correlated (p < 0.001) to number of moths (N), while minimum temperature (T min ), minimum relative humidity (RH min ), mean relative humidity (RH mean ), maximum relative humidity (RH max ) and host plant (h) were negatively correlated (p < 0.001) to N. Using Waikato Environment for Knowledge Analysis (WEKA) data-mining techniques, two models were developed: (a) M5P decision-tree algorithm associated with nine linear models (LMs) and (b) principal component analysis (PCA) based on four principal components. Both approaches identified M as the major predictor of moth abundance, followed by h and farming region (R). However, R was a function of T max (positive auto-correlation) and RH max (negative auto-correlation). These results provide simplified relative contribution of heterogeneous factors in influencing P. xylostella spatio-temporal abundance, essential for early warning systems in pest management. This is an important component of sustainable pest management aimed at managing insect pests and minimizing pesticides abuse in brassica production systems. K E Y W O R D S hyperparameter optimization, insect population dynamics, integrated pest management, predictive models, small-scale farmers, Waikato Environment for Knowledge Analysis (WEKA) | 671 MACHEKANO Et Al.

show abstract

“…However, when dealing with large datasets, decision trees tend to become very large, complex, and difficult to understand. Consequently, the rules constructed through 'Divide and Conquer' strategy and extracted from the resulting decision tree inherit the tree's complexity and thus may have unnecessary repeated tests which can lead to redundant rulesets [1], [2]. In this context, research has been carried out aiming to overcome or reduce these problems and to produce a simple reliable ruleset using pruning methods.…”

Section: Introductionmentioning

confidence: 99%

“…In this context, research has been carried out aiming to overcome or reduce these problems and to produce a simple reliable ruleset using pruning methods. Among others, C4.5rules [3], [4] and CART [2] are examples of such algorithms. Nevertheless, [5]- [7] argue that there is no single study which adequately achieves this goal.…”

Section: Introductionmentioning

confidence: 99%

A Rule-Based Classifier with Accurate and Fast Rule Term Induction for Continuous Attributes

Almutairi

Stahl

Bramer³

2018

2018 17th IEEE International Conference on Machine Learning and Applications (ICMLA)

View full text Add to dashboard Cite

Rule-based classifiers are considered more expressive, human readable and less prone to over-fitting compared with decision trees, especially when there is noise in the data. Furthermore, rule-based classifiers do not suffer from the replicated subtree problem as classifiers induced by top down induction of decision trees (also known as 'Divide and Conquer'). This research explores some recent developments of a family of rulebased classifiers, the Prism family and more particular G-Prism-FB and G-Prism-DB algorithms, in terms of local discretisation methods used to induce rule terms for continuous data. The paper then proposes a new algorithm of the Prism family based on a combination of Gauss Probability Density Distribution (GPDD), InterQuartile Range (IQR) and data transformation methods. This new rule-based algorithm, termed G-Rules-IQR, is evaluated empirically and outperforms other members of the Prism family in execution time, accuracy and tentative accuracy.

show abstract

Techniques of Decision Tree Induction

Cited by 9 publications

References 127 publications

Multirate Processing with Selective Subbands and Machine Learning for Efficient Arrhythmia Classification

Multirate Processing with Selective Subbands and Machine Learning for Efficient Arrhythmia Classification

Disentangling factors limiting diamondback moth, Plutella xylostella (L.), spatio‐temporal population abundance: A tool for pest forecasting

A Rule-Based Classifier with Accurate and Fast Rule Term Induction for Continuous Attributes

Contact Info

Product

Resources

About