A Business Rule for a B-School using Machine Learning

Ganatra, Dhimant

doi:10.30534/ijatcse/2019/145862019

Cited by 3 publications

(8 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The objective here is to improve the accuracy especially of the positive class of the classifier model developed by [1]. The classifier model was built following the requirement of the institute to differentiate applicants into two categoriesplaceable and non-placeable.…”

Section: Methodsmentioning

confidence: 99%

“…As per [1], 215 students who completed their MBA from the business school have been selected for the study. Here the response variable is a two-class categorical variable -Placement with labels as Placed and Not Placed.…”

Section: Methodsmentioning

confidence: 99%

“…A business rule has been developed by [1] for a B-School using which it was possible to divide applicants for admission into two classes -Placeable and Not Placeable. Owing to the impact placement has on a B-School's brand image, the management wanted to devise a strategy which can help differentiate the pool of applicants.…”

Section: The Classification Settingmentioning

confidence: 99%

“…The admission team was keen to use the power of analytics in arriving at this decision and also wanted to keep the model simple and interpretable. Given this context, using the classification tree algorithm, a model was developed by [1]. On account of the recent developments across the world owing to the Covid-19 pandemic, there has been a spurt in the number of applicants for admission.…”

Section: The Classification Settingmentioning

confidence: 99%

See 3 more Smart Citations

Improving classification accuracy :The KNN approach

Ganatra¹

2020

IJATCSE

View full text Add to dashboard Cite

The non-parametric tree-based methods are the go-to choice in a classification setting. They are simple and very useful for interpretation especially in contexts that require a business rule. They are however not very competitive in terms of predictive power when compared to other supervised learning approaches. Combing multiple trees in order to improve prediction accuracy is the next best option. Ensemble methods are powerful prediction models but they come at a cost. The loss of interpretability on account of aggregating trees may not be feasible in every decision making scenario. Also, depending on the business goal, the class-specific performance may be crucial. The true positive rate and positive predicted value may be more important than the overall accuracy.A non-parametric approach like K-Nearest Neighbours (KNN) can be superior when we have a complicated decision boundary,. Though simple, KNN often produces a classifier that is closer to the optimal Bayes classifier. For the B-School in question which needs to classify applicants into placeableor nonplaceablebased on their past academic performance, a comparison of both the approaches is made to identify a superior performer on the positive class. A model's sensitivity is more crucial than the reduction in the overall error in the given scenario.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: The Classification Settingmentioning

confidence: 99%

Section: The Classification Settingmentioning

confidence: 99%

See 2 more Smart Citations

Improving classification accuracy :The KNN approach

Ganatra¹

2020

IJATCSE

View full text Add to dashboard Cite

show abstract

“…A large number of works are devoted to the problem of choosing the control of learning [1]- [7]. These works are related both to the search for technical teaching aids and modeling of the learning process.…”

Section: Introductionmentioning

confidence: 99%

Model of Accumulation and Loss of Knowledge in Computerization Systems of Education with Remote Access

Omarov¹

2020

IJETER

View full text Add to dashboard Cite

The article discusses the approach to the methods of creating expert assessments to determine the level of knowledge of students in education computerization systems with remote access. This is achieved through the use of the probabilistic approach and Bayesian processes in comparison with the approach based on the theory of fuzzy sets. A model for the accumulation and loss of knowledge is proposed. A life-cycle model of a student's knowledge system has been developed. The model of finding the optimal periodicity of the learning process with forced interruptions is investigated.

show abstract

The Sensitivity Conundrum – Random Forest or Boosting

Ganatra¹

2020

IJETER

Self Cite

View full text Add to dashboard Cite

In the classification context, tree-based models are simple and useful for interpretation. However when it comes to model accuracy the single-tree model does not match the power of other supervised learning approaches. By aggregating trees a model's accuracy can be improved. Ensemble methods like random forest and boosting combine predictions from multiple models into one that is far superior to the individual models. Depending on the business goal, the accuracy paradox may come into play. The class statistics, precision and recall may be more important than the overall accuracy. The True Positive Rate varies based on the type of ensemble used among other factors. While both random forest and boosting lead to some loss of interpretability, the improvement in sensitivity will outweigh this loss. By tuning several of the model's parameters it is possible to achieve higher levels on any of the four counts in the confusion matrix. While both random forest and boosting use trees as base learners, they differ primarily in the way the trees are built. A comparison of both the approaches is made to identify a superior performer on the positive class.

show abstract

A Business Rule for a B-School using Machine Learning

Cited by 3 publications

References 9 publications

Improving classification accuracy :The KNN approach

Improving classification accuracy :The KNN approach

Model of Accumulation and Loss of Knowledge in Computerization Systems of Education with Remote Access

The Sensitivity Conundrum – Random Forest or Boosting

Contact Info

Product

Resources

About