Software Bug Prediction Prototype Using Bayesian Network Classifier: A Comprehensive Model

Software bugs (or malfunctions) pose a serious threat to software developers with many known and unknown bugs that may be vulnerable to computer systems, demanding new methods, analysis, and techniques for efficient bug detection and repair of new unseen programs at a later stage. This chapter uses evolutionary grey wolf (GW) search optimization as a feature selection technique to improve classifier efficiency. It is also envisaged that software error detection would consider the nature of the error when repairing it for remedial action instead of simply finding it either faulty or non-defective. To address this problem, the authors use bug severity multi-class classification to build an efficient and robust prediction model using multilayer perceptron (MLP), logistic regression (LR), and random forest (RF) for bug severity classification. Both tests are performed on two software error datasets, namely Ant 1.7 and Tomcat.

Download Full-text

Human Arthritis Analysis in Fog Computing Environment Using Bayesian Network Classifier and Thread Protocol

IEEE Consumer Electronics Magazine ◽

10.1109/mce.2019.2941456 ◽

2020 ◽

Vol 9 (1) ◽

pp. 88-94 ◽

Cited By ~ 27

Author(s):

Sudeep Tanwar ◽

Jayneel Vora ◽

Shriya Kaneriya ◽

Sudhanshu Tyagi ◽

Neeraj Kumar ◽

...

Keyword(s):

Bayesian Network ◽

Fog Computing ◽

Computing Environment ◽

Bayesian Network Classifier

Download Full-text

Learning semi-lazy Bayesian network classifier under the c.i.i.d assumption

Knowledge-Based Systems ◽

10.1016/j.knosys.2020.106422 ◽

2020 ◽

Vol 208 ◽

pp. 106422 ◽

Cited By ~ 3

Author(s):

Yang Liu ◽

Limin Wang ◽

Musa Mammadov

Keyword(s):

Bayesian Network ◽

Bayesian Network Classifier

Download Full-text

A Remote Sensing-Based Application of Bayesian Networks for Epithermal Gold Potential Mapping in Ahar-Arasbaran Area, NW Iran

Remote Sensing ◽

10.3390/rs12010105 ◽

2019 ◽

Vol 12 (1) ◽

pp. 105 ◽

Cited By ~ 9

Author(s):

Seyed Mohammad Bolouki ◽

Hamid Reza Ramazi ◽

Abbas Maghsoudi ◽

Amin Beiranvand Pour ◽

Ghahraman Sohrabi

Keyword(s):

Remote Sensing ◽

Bayesian Network ◽

Hydrothermal Alteration ◽

Gold Mineralization ◽

Epithermal Gold ◽

Alteration Zones ◽

Mineral Potential ◽

Potential Map ◽

Bayesian Network Classifier ◽

Hydrothermal Alteration Zones

Mapping hydrothermal alteration minerals using multispectral remote sensing satellite imagery provides vital information for the exploration of porphyry and epithermal ore mineralizations. The Ahar-Arasbaran region, NW Iran, contains a variety of porphyry, skarn and epithermal ore deposits. Gold mineralization occurs in the form of epithermal veins and veinlets, which is associated with hydrothermal alteration zones. Thus, the identification of hydrothermal alteration zones is one of the key indicators for targeting new prospective zones of epithermal gold mineralization in the Ahar-Arasbaran region. In this study, Landsat Enhanced Thematic Mapper+ (Landsat-7 ETM+), Landsat-8 and Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) multispectral remote sensing datasets were processed to detect hydrothermal alteration zones associated with epithermal gold mineralization in the Ahar-Arasbaran region. Band ratio techniques and principal component analysis (PCA) were applied on Landsat-7 ETM+ and Landsat-8 data to map hydrothermal alteration zones. Advanced argillic, argillic-phyllic, propylitic and hydrous silica alteration zones were detected and discriminated by implementing band ratio, relative absorption band depth (RBD) and selective PCA to ASTER data. Subsequently, the Bayesian network classifier was used to synthesize the thematic layers of hydrothermal alteration zones. A mineral potential map was generated by the Bayesian network classifier, which shows several new prospective zones of epithermal gold mineralization in the Ahar-Arasbaran region. Besides, comprehensive field surveying and laboratory analysis were conducted to verify the remote sensing results and mineral potential map produced by the Bayesian network classifier. A good rate of agreement with field and laboratory data is achieved for remote sensing results and consequential mineral potential map. It is recommended that the Bayesian network classifier can be broadly used as a valuable model for fusing multi-sensor remote sensing results to generate mineral potential map for reconnaissance stages of epithermal gold exploration in the Ahar-Arasbaran region and other analogous metallogenic provinces around the world.

Download Full-text