Detection of nicotine content impact in tobacco manufacturing using computational intelligence.

A study is presented for the detection of nicotine impact in different cigarette type, using recorded data and Computational Intelligence techniques. Recorded puffs are processed using Continuous Wavelet Transform and used to extract time-frequency features for normal and abnormal puffs conditions. The wavelet energy distributions are used as inputs to classifiers based on Adaptive Neuro-Fuzzy Inference Systems (ANFIS) and Genetic Algorithms (GAs). The number and the parameters of Membership Functions are used in ANFIS along with the features from wavelet energy distributionare selected using GAs, maximising the diagnosis success. GA with ANFIS (GANFIS) are trained with a subset of data with known nicotine conditions. The trained GANFIS are tested using the other set of data (testing data). A classical method by High-Performance Liquid Chromatography is also introduced to solve this problem, respectively. The results as well as the performances of these two approaches are compared. A combination of these two algorithms is also suggested to improve the efficiency of this solution procedure. Computational results show that this combined algorithm is promising.

Optimized detection of tar content in the manufacturing process using adaptive neuro-fuzzy inference systems.

The purpose of this study is to model and optimize the detection of tar in cigarettes during the manufacturing process and show that low yield cigarettes contain similar levels of nicotine as compared to high yield cigarettes while B (Benzene), T(toluene) and X (xylene) (BTX) levels increase with increasing tar yields. A neuro-fuzzy system which comprises a fuzzy inference structure is used to model such a system. Given a training set of samples, the Adaptive Neuro-Fuzzy Inference System (ANFIS) classifiers learned how to differentiate a new case in the domain. The ANFIS classifiers were used to detect the tar in smoke condensate when five basic features defining cigarette classes indications were used as inputs. A classical method by High Performance Liquid Chromatography (HPLC) is also introduced to solve this problem. At last the performances of these two methods are compared.

Object-oriented integrated approach for the design of scalable ECG systems.

The paper presents the implementation of Object-Oriented (OO) integrated approaches to the design of scalable Electro-Cardio-Graph (ECG) Systems. The purpose of this methodology is to preserve real-world structure and relations with the aim to minimize the information loss during the process of modeling, especially for Real-Time (RT) systems. We report on a case study of the design that uses the integration of OO and RT methods and the Unified Modeling Language (UML) standard notation. OO methods identify objects in the real-world domain and use them as fundamental building blocks for the software system. The gained experience based on the strongly defined semantics of the object model is discussed and related problems are analyzed.

Artificial Intelligence in Prediction of Secondary Protein Structure Using CB513 Database.

In this paper we describe CB513 a non-redundant dataset, suitable for development of algorithms for prediction of secondary protein structure. A program was made in Borland Delphi for transforming data from our dataset to make it suitable for learning of neural network for prediction of secondary protein structure implemented in MATLAB Neural-Network Toolbox. Learning (training and testing) of neural network is researched with different sizes of windows, different number of neurons in the hidden layer and different number of training epochs, while using dataset CB513.