Allan Melvin Andrew
Thumbnail Image
Preferred name
Allan Melvin Andrew
Official Name
Allan Melvin, Andrew
Alternative Name
Andrew, Allan Melvin
Andrew, Allan M.
Melvin, A. A.
Main Affiliation
Scopus Author ID
36469871200
Researcher ID
GDH-3820-2022

Research Output

Filters

2
2
Reset filters

Settings
Now showing 1 - 2 of 2
  • Publication
    Intelligent classifier for incipient phase fire in building
    ( 2017)
    Allan Melvin Andrew
    Early fire detection is one of the most promising sub-fields in indoor air quality research. Ability to give early fire indication can help the building occupants to take responsive actions in order to prevent the fire. Delay in having such indication not only leading to property and money losses, but also life losses. This research is a preliminary research intended to detect the early fire and the material (common fire sources and building construction materials) involved in the fire using intelligent classifier. Indoor Air Quality (IAQ) database is formed as the testing database, while Portable Electronic Nose 3 (PEN3) database is formed to verify the IAQ database. The databases consist of gas sensor inputs from the test materials, heated up at different temperatures in the testbed. Seven temperatures, range from 50°C up to 250°C have been tested. For incipient phase fire, data for temperature range 75°C up to 125°C shows a very significant result. The data is pre-processed and normalised into five types of normalised features. Out of the five normalised features, only three were statistically selected for proposed multi- stage feature selection and feature fusion process. As an output to the proposed process, a new robust feature, IAQ-Hybrid feature is formed. IAQ-Hybrid feature is consisting of dimensionally reduced principal components fused by the feature fusion technique. ANOVA F- Test and Principal Component Analysis are used for selecting the useful and non- redundant data for the proposed feature formulation. The proposed feature and the other normalised features (three types of normalised features which were statistically selected earlier) are tested with various common unsupervised, semi- supervised and supervised classifiers.
      1  6
  • Publication
    Multi-Stage feature selection based intelligent classifier for classification of incipient stage fire in building
    ( 2016)
    Allan Melvin Andrew
    ;
    Ammar Zakaria
    ;
    Shaharil Mad Saad
    ;
    Ali Yeon Md Shakaff
    In this study, an early fire detection algorithm has been proposed based on low cost array sensing system, utilising off- the shelf gas sensors, dust particles and ambient sensors such as temperature and humidity sensor. The odour or “smellprint” emanated from various fire sources and building construction materials at early stage are measured. For this purpose, odour profile data from five common fire sources and three common building construction materials were used to develop the classification model. Normalised feature extractions of the smell print data were performed before subjected to prediction classifier. These features represent the odour signals in the time domain. The obtained features undergo the proposed multi-stage feature selection technique and lastly, further reduced by Principal Component Analysis (PCA), a dimension reduction technique. The hybrid PCA-PNN based approach has been applied on different datasets from in-house developed system and the portable electronic nose unit. Experimental classification results show that the dimension reduction process performed by PCA has improved the classification accuracy and provided high reliability, regardless of ambient temperature and humidity variation, baseline sensor drift, the different gas concentration level and exposure towards different heating temperature range.
      1  12