Regular paperLearning using an artificial immune system
Abstract
In this paper we describe an artificial immune system (AIS) which is based upon models of the natural immune system. This natural system is an example of an evolutionary learning mechanism which possesses a content addressable memory and the ability to «forget» little-used information. It is also an example of an adaptive non-linear network in which control is decentralized and problem processing is efficient and effective. As such, the immune system has the potential to offer novel problem solving methods. The AIS is an example of a system developed around the current understanding of the immune system. It illustrates how an artificial immune system can capture the basic elements of the immune system and exhibit some of its chief characteristics. We illustrate the potential of the AIS on a simple pattern recognition problem. We then apply the AIS to a real-world problem: the recognition of promoters in DNA sequences. The results obtained are consistent with other appproaches, such as neural networks and Quinlan's ID3 and are better than the nearest neighbour algorithm. The primary advantages of the AIS are that it only requires positive examples, and the patterns it has learnt can be explicitly examined. In addition, because it is self-organizing, it does not require effort to optimize any system parameters.
References (0)
Cited by (320)
IM-NKA: A Natural Killer cell Algorithm for earthquake prediction based on extremely imbalanced precursor data
2022, Applied Soft ComputingEarthquake prediction plays a vital role in reducing the risk of earthquakes to human beings. However, the precursor seismic data is extremely imbalanced, resulting in unsatisfactory performances of standard algorithms on earthquake prediction. In this paper, we propose a Natural Killer cell Algorithm inspired by the Induction and phenotype discrimination Mechanism of natural killer cells (IM-NKA) for earthquake prediction. IM-NKA first generates induction pathogens (synthetic samples) in the hyper-sphere area around each pathogen (selected minority instance), called NKI. After balancing the dataset, the IM-NKA generates the phenotype detectors. Then, the IM-NKA is used to predict earthquakes with a magnitude larger than 3.5 in the next day. The experiment consists of two parts: the first part compares the NKI with four oversampling methods and demonstrates that using NKI as an oversampling method can effectively improve the quality of synthetic samples for unbalanced earthquake prediction; the second part compares the performance of IM-NKA with eight different classifiers. The experiment results show that the performance of earthquake prediction using IM-NKA also significantly outperforms other classifiers.
A comprehensive survey on recent metaheuristics for feature selection
2022, NeurocomputingFeature selection has become an indispensable machine learning process for data preprocessing due to the ever-increasing sizes in actual data. There have been many solution methods proposed for feature selection since the 1970s. For the last two decades, we have witnessed the superiority of metaheuristic feature selection algorithms, and tens of new ones are being proposed every year. This survey focuses on the most outstanding recent metaheuristic feature selection algorithms of the last two decades in terms of their performance in exploration/exploitation operators, selection methods, transfer functions, fitness value evaluations, and parameter setting techniques. Current challenges of the metaheuristic feature selection algorithms and possible future research topics are examined and brought to the attention of the researchers as well.
A survey of recommender systems with multi-objective optimization
2022, NeurocomputingRecommender systems have been widely applied to several domains and applications to assist decision making by recommending items tailored to user preferences. One of the popular recommendation algorithms is the model-based approach which optimizes a specific objective to improve the recommendation performance. These traditional recommendation models usually deal with a single objective, such as minimizing the prediction errors or maximizing the ranking quality of the recommendations. In recent years, there is an emerging demand for multi-objective recommender systems in which multiple objectives are considered and the recommendations can be optimized by the multi-objective optimization. For example, a recommendation model may be built by optimizing multiple metrics, such as accuracy, novelty and diversity of the recommendations. The multi-objective optimization methodologies have been well developed and applied to the area of recommender systems. In this article, we provide a comprehensive literature review of the multi-objective recommender systems. Particularly, we identify the circumstances in which a multi-objective recommender system could be useful, summarize the methodologies and evaluation approaches in these systems, point out existing challenges or weaknesses, finally provide the guidelines and suggestions for the development of multi-objective recommender systems.
Adaptation of the clonal selection algorithm to the real-time coordinated Volt/VAr control through a software-in-the-loop strategy
2021, Electric Power Systems ResearchWith the advent of smart grid technologies, modern Distribution Management Systems (DMSs) enable the real-time operation of advanced automation functions such as Coordinated Volt/VAr Control (CVVC). Typically, control rule-based techniques have been proposed to efficiently carry out this important function in Distribution Networks (DNs). In this paper, a novel adaptation of the clonal selection theory of immune systems (CLONALG - Clonal Selection Algorithm) is proposed for real-time optimal solution of CVVC problem in radial DNs composed of multiple control equipment. To do this, a Software-in-the-Loop (SIL) strategy is developed, through the use of the Real-Time Digital Simulator (RTDS) and Matlab® program. Implementation details of the referred strategy and the corresponding adaptation of the CLONALG-based CVVC are discussed in this paper. Simulations results demonstrated that the proposed adaptation of the CLONALG technique is suitable to solve the CVVC problem in a maximum time of less than 10 min, in accordance with the Brazilian electric utilities’ need. On the other hand, the developed SIL strategy proved to be effective to analyzing the time domain behavior of CVVC algorithms, designed to operate in real time in a modern DMS.
Continual learning classification method with new labeled data based on the artificial immune system
2020, Applied Soft Computing JournalIn this paper, a new supervised learning classification method, continual learning classification method with new labeled data based on the artificial immune system (CLCMNLD), is proposed as a new way to improve the classification performance in real-time by continually learning the new labeled data during the testing stage. It is inspired by the mechanism that vaccines can enhance immunity. New types of memory cells were continuously cultured by learning new labeled data during the testing stage. CLCMNLD will degenerate into a common supervised learning classification method when there is no new labeled data comes out during the testing stage. The effectiveness of the proposed CLCMNLD is tested on twenty well-known datasets from the UCI Machine Learning Repository that are commonly used in the domain of data classification. The experiments reveal that CLCMNLD has better classification performance when it degenerates into a common supervised learning classification method, and it outperforms the other methods when there are some new labeled data comes out during the testing stage. The more types of new labeled data, the more advantages it has.
Data-driven modeling and learning in science and engineering
2019, Comptes Rendus - MecaniqueIn the past, data in which science and engineering is based, was scarce and frequently obtained by experiments proposed to verify a given hypothesis. Each experiment was able to yield only very limited data. Today, data is abundant and abundantly collected in each single experiment at a very small cost. Data-driven modeling and scientific discovery is a change of paradigm on how many problems, both in science and engineering, are addressed. Some scientific fields have been using artificial intelligence for some time due to the inherent difficulty in obtaining laws and equations to describe some phenomena. However, today data-driven approaches are also flooding fields like mechanics and materials science, where the traditional approach seemed to be highly satisfactory. In this paper we review the application of data-driven modeling and model learning procedures to different fields in science and engineering.