Open AccessTPR, PPV and ROC based Performance Measurement and Optimization of Human Face Recognition of IoT Enabled Physical Location Monitoring
Author(s) -
Ajitkumar Shitole,
Manoj Devare
Publication year - 2019
Publication title -
international journal of recent technology and engineering
Language(s) - English
Resource type - Journals
ISSN - 2277-3878
DOI - 10.35940/ijrte.b3186.078219
Subject(s) - naive bayes classifier , computer science , artificial intelligence , classifier (uml) , decision tree , facial recognition system , machine learning , face (sociological concept) , receiver operating characteristic , pattern recognition (psychology) , internet of things , logistic regression , data mining , support vector machine , social science , sociology , embedded system
This paper describes the construction of Internet of Things (IoT) enabled system which not only captures the sensors data in textual and numeric form but also performs live human face recognition to monitor physical location effectively. The dataset used in order to apply supervised machine learning algorithms is the combination of automatically captured live sensor data along with name of the human face recognized or unknown and additional manually introduced class label. Performance measurement of face recognition is done with the help of Decision Tree (DT), K-Nearest Neighbors (KNN), Naïve Bayes (NB) and Logistic Regression (LR). The results show that DT gives the best performance with respect to classifier’s accuracy; True Positive Rate, Positive Predictive Value and area under curve of Receiver Operating Characteristics (ROC) for face recognition prediction whether the recognized face is true or false.