Fuzzy Logic

Inhaltsverzeichnis

1. Frontmatter

2. Fuzzy Logic, a Logician’s Perspective

   Patrick Fogarty

   Abstract

   Fuzzy logic arises from an attempt to manage the inherent vagueness there is in the language we use when discussing our world—it is a formal treatment of vague predicates. This chapter will describe how this formal structure has come about, from origins in philosophical thought, through the development of non-standard logics. It will explore, from a logician’s perspective, useful tools using fuzzy set theories, such as Basic Fuzzy Logic (BL) and T-Norm Fuzzy logics, deployed in computer systems today. It is intended to detail the techniques used to set up such theories and to review the relationship that Logic bears to them. In conclusion, it is proposed that further suggested theoretical investigations might yield useful practical results.

3. A Fuzzy Approach to Sentiment Analysis at the Sentence Level

   Orestes Appel, Francisco Chiclana, Jennifer Carter, Hamido Fujita

   Abstract

   The objective of this chapter is to present a hybrid approach to the Sentiment Analysis problem focused on sentences or snippets. This new method is centred around a sentiment lexicon enhanced with the assistance of SentiWordNet and fuzzy sets to estimate the semantic orientation polarity and intensity for sentences. This provides a foundation for computing with sentiments. The proposed hybrid method is applied to three different datasets and the results achieved are compared to those obtained using Naïve Bayes (NB) and Maximum Entropy (ME) techniques. It is demonstrated through experimentation that this hybrid approach is more accurate and precise than both NB and ME techniques. Furthermore, it is shown that when applied to datasets containing snippets, the proposed method performs similar to state-of-the-art techniques.

4. Consensus in Sentiment Analysis

   Orestes Appel, Francisco Chiclana, Jennifer Carter, Hamido Fujita

   Abstract

   The objective of this chapter is to present a method applicable in group decision-making where computing the opinion of the majority of participants is key. In this article, we present a method that makes use of Induced Ordered Weighted Averaging (IOWA) operators to aggregate a majority opinion out of a number of Sentiment Analysis (SA) classification systems. The numerical output of each SA classification method is used as input to a carefully chosen IOWA operator that is semantically equivalent to the fuzzy linguistic quantifier ‘most of’. The object of the aggregation will be the intensity of the previously determined sentence polarity in such a way that the results represent what the majority thinks.

5. Fostering Positive Personalisation Through Fuzzy Clustering

   Raymond Moodley

   Abstract

   Elements of personalisation theory, personalisation engine modelling, and artificial intelligence algorithms using Fuzzy clustering are combined to provide a useful approach to enable positive personalisation that produces valuable outcomes for both the user and organisation. A recent case study in grocery retail, which applied this approach, shows that it is possible for a store to offer personalised promotions to its customers and in the process increase its market share and increase savings for customers. Adopting this personalisation approach creates a mutually beneficial environment, which is the essence of fostering positive personalisation.

6. Diagnosing Alzheimer’s Disease Using a Self-organising Fuzzy Classifier

   Jonathan Stirling, Tianhua Chen, Magda Bucholc

   Abstract

   Dementia is one of the major causes of disability and dependency among older people worldwide. Without treatment currently available to cure dementia or to alter its progressive course, one of the principal goals for dementia care set by the World Health Organization is the early diagnosis in order to promote early and optimal management. In recognition of the potentials of fuzzy systems in effectively dealing with medical data, this chapter investigates the use of a very recently proposed Self-Organising Fuzzy (SOF) classifier for the prediction of Alzheimer’s Disease against Mild Cognitive Impairment and being Cognitively Unimpaired with patient observations provided by the renowned Alzheimer’s Disease Neuroimaging Initiative repository. The experimental study demonstrates the effectiveness of SOF, especially in combined use with the Recursive Feature Elimination feature selection.

7. Autism Spectrum Disorder Classification Using a Self-organising Fuzzy Classifier

   Jonathan Stirling, Tianhua Chen, Marios Adamou

   Abstract

   Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder that covers a range of symptoms such as impaired social skills and repetitive behaviours. The diagnosis of ASD in clinics is typically lengthy and cost-ineffective. Recent advances in machine learning could facilitate more efficient and effective detection of ASD. However, fuzzy systems, as a significant soft computing technique, have been sporadically applied in the diagnosis of ASD. This chapter, therefore, examines the use of a recently proposed self-organising fuzzy classifier with application to the “autism screening adult” data retrieved from a mobile application.

8. An Outlier Detection Informed Aggregation Approach for Group Decision-Making

   Chunru Chen, Tianhua Chen, Zhongmin Wang, Yanping Chen, Hengshan Zhang

   Abstract

   In group decision-making, owing to differences that may result from perspectives such as experience and knowledge, the evaluations about the same decision problem provided by different crowd participants may have great differences. Those with huge differences in evaluations from most participants are termed outliers in this chapter. Reaching a decision consensus that satisfies most people is very difficult. In order to solve this problem, many researchers have conducted consensus research. To avoid this problem, this chapter proposes an expert opinions aggregation method based on outlier detection. First, the decision-maker evaluates the decision problem based on the Pythagorean Fuzzy Sets (PFSs) from the positive and the negative views. Second, the outliers of expert opinions are detected and then aggregated to obtain the overall decision result. The effectiveness of the proposed method is finally demonstrated using a case study.

9. Novel Aggregation Functions Based on Domain Partition with Concentrate Region of Data

   Hengshan Zhang, Tianhua Chen, Zhongmin Wang, Yanpin Chen

   Abstract

   Combining numerous input arguments, specially in case most arguments lie in a concentrate region, is a complex issue. This chapter proposes to partition the input domain on the basis of the concentrate region, which can then be tackled based on the sub-regions. Furthermore, two bi-variate aggregation functions are proposed, which aim to behave differently in response to the corresponding sub-regions. The bi-variate functions are extended further into multivariate functions in combination with the popular Ordered Weighted Averaging OWA operators. Finally, the proposed aggregation functions are assessed using a case study where the maintainability of the Linux Kernels is evaluated, demonstrating the effectiveness of the proposed functions.

10. Applying Fuzzy Pattern Trees for the Assessment of Corneal Nerve Tortuosity

    Pan Su, Xuanhao Zhang, Hao Qiu, Jianyang Xie, Yitian Zhao, Jiang Liu, Tianhua Chen

    Abstract

    The tortuosity of corneal nerve fibers is correlated with a number of diseases such as diabetic neuropathy. The assessment of corneal nerve tortuosity level in in vivo confocal microscopy (IVCM) images can inform the detection of early diseases and further complications. With the aim to assess the corneal nerve tortuosity accurately as well as to extract knowledge meaningful to ophthalmologists, this chapter proposes a fuzzy pattern tree-based approach for the automated grading of corneal nerves’ tortuosity based on IVCM images. The proposed method starts with the deep learning-based image segmentation of corneal nerves and then extracts several morphological tortuosity measurements as features for further processing. Finally, the fuzzy pattern trees are constructed based on the extracted features for the tortuosity grading. Experimental results on a public corneal nerve data set demonstrate the effectiveness of fuzzy pattern tree in IVCM image tortuosity assessment.

11. A Mamdani Fuzzy Logic Inference System to Estimate Project Cost

    Daniel Helder Maia, Arjab Singh Khuman

    Abstract

    The precision and reliability of estimations of project costs are essential, especially in significant cooperation. The level of uncertainty when estimating projects can cause issues down the line during a project. For generations, humans are more often than always in a predicament where estimation for a project size or cost appears to be complicated. The methodology adopted in this research included using the literature to review the topic of project estimation and explore the use of fuzzy logic in order to define an initial fuzzy system. The development of a system to estimate project costs is based on findings from the literature. This work seeks to demonstrate the benefits of using fuzzy logic in estimating the cost for business. Analysis of the results attained during testing and research shows that the system could be beneficial for estimating the cost of projects. The results show that the system can produce an appropriate result when estimating project cost. The study concludes that there is still room for improvement and that further development and testing could lead to improvements; however, the current system gives a foundation for further development such that the system can be put to use in a real-world situation. Whether it is for business or personal circumstances where any or most cases, cost estimation is required.

12. Artificial Intelligence in FPS Games: NPC Difficulty Effects on Gameplay

    Adam Hubble, Jack Moorin, Arjab Singh Khuman

    Abstract

    This report explores the use of fuzzy logic within computer games, with specific respect to their use of Artificial Intelligence (AI) within the games’ enemy Non-Player Characters (NPCs), in order to affect the game’s overall difficulty. The way in which AI is affected varies across different games; games within the same genre often share multiple statistics and values, and these can be applied to an NPC in order to make the game easier or harder. Games within the First-Person Shooter (FPS) genre, for example, can always affect their difficulty by changing an enemy character’s accuracy with weapons or overall damage output as these would all change how likely they are to defeat the player in a combat scenario. In this document, we will be detailing the development and structure of the multiple input Mamdani styled fuzzy inference system (FIS) that we developed in order to rate a given NPC’s difficulty based on the rankings they have been given for these shared statistics.

13. Adaptive Cruise Control Using Fuzzy Logic

    Nathan Lloyd, Arjab Singh Khuman

    Abstract

    Modern transportation undoubtedly provides a plethora of beneficial qualities; qualities that not only dramatically improve the efficiency and speed of travel, but also provide materialistic comforts for the inhabitants of the vehicle. Whilst these advancements have generally improved the quality of life for users, it begs the question: can modern technologies be utilized to augment vehicles further? This chapter will engage intelligent transportation systems (ITS), specifically automatic cruise control (ACC) and the utilization of fuzzy inference systems (FIS), analyzing their successful implementation, posing a bespoke system and how the ITS field can be improved further.

14. Automatic Camera Flash Using a Mamdani Type One Fuzzy Inference System

    Sophie Hughes, Arjab Singh Khuman

    Abstract

    Photography is an enjoyable hobby for many people, with many systems having been developed to make it easier for newcomers to begin learning how to take a quality photograph. Features such as automatic aperture and shutter speed allow the user to take a photo without any prior knowledge as to how these two should be manipulated in order to take a good photo. However, a feature that has not currently been explored is an automatic camera flash that will change intensity based on a number of factors, as current automatic flash systems will simply either activate a flash or not based on the perceived light levels of the image. This chapter will utilise a Mamdani type one fuzzy inference system in order to demonstrate how an automatic camera flash could potentially work, justifying each input used as well as discussing any limitations and possible improvements.

15. The Application of Fuzzy Logic in Determining Outcomes of eSports Events

    Spencer Deane, Arjab Singh Khuman

    Abstract

    As eSports skyrocket in popularity, the saturation of top talent intensifies. Hundreds of millions of dollars in prize money are distributed amongst this talent, resulting in fierce competition. To get ahead, players go to extreme measures to gain marginal performance increases. Besides intense training and performance enhancing drugs, athletes seek intelligent analytical tools which can provide useful insights into a player’s strengths and weaknesses. This report showcases a fuzzy system which uses real-world data and determines a player’s percentage chance of winning a duel in the online first-person shooter video game Counter-Strike Global Offensive, one of the leading eSports.

16. Water Carbonation Fuzzy Inference System

    William Chapman, Arjab Singh Khuman

    Abstract

    This report will be looking at how fuzzy logic is used to create a system which automatically carbonates water to create sparkling water. This is a topic that has not been discussed a large amount and there is little about it in the associated literature. The system created in this work uses the research available to create a system that carbonates water based on the temperature of the water, the amount of water being carbonated and the sparkling preference of the user. This is then processed in the Water Carbonation Fuzzy Inferencing System (FIS) which outputs to give the Carbon Dioxide Usage. This system is made for domestic water carbonation products and could be extended to larger or smaller products of the type. Several tests have been carried out to measure the success of the system. Changes are then made, and the system is tested again to make sure that the system has been improved. Tests are continued until the system is efficient for the purpose and all the different possible parameters are checked. A critical reflection on the work illuminates the good points, limitations and where improvements could be made.

17. Correction to: An Outlier Detection Informed Aggregation Approach for Group Decision-Making

    Chunru Chen, Tianhua Chen, Zhongmin Wang, Yanping Chen, Hengshan Zhang