International Conference on Information and Communication Technology and Its Applications (ICTA 2016) Federal University of Technology, Minna, Nigeria November 28 – 30, 2016 Enhanced Select and Test (eST) Algorithm: Framework for Diagnosing and Monitoring Related Ailments Olaide Nathaniel Oyelade1, Enesi Femi Aminu2, Solomon Adelowo Adepoju3, and Ibrahim Shehi Shehu4 1 Department of Computer Science, Ahmadu Bello University, Zaria, Nigeria 2,3,4 Department of Computer Science, Federal University of Technology, Minna, Nigeria 1 onoyelade@abu.edu.ng, {2enesifa, 3solomon.adepoju, 4ibrahim.shehu} @futminna.edu.ng Abstract—Diagnosis, prediction, machine learning, and reasoning algorithm is an exceptional tool in incorporating decision making are all areas of application of artificial autonomous (health) systems. intelligence. Particularly, intelligent (medical) diagnosis This research in progress argues that employing the use of systems are now becoming pervasive providing support to some semantic web technologies in ST algorithm will yield healthcare delivery. However, there is a lack of precision and a higher precision, and an inference making medical approximation of the algorithms driving such diagnostics diagnostic reasoning system. systems. Though there is a number of reasoning algorithms for carrying out this diagnostic task, the precision of these diagnostic algorithms are being impaired by their reasoning II. RELATED WORKS structures. This paper reviews and provides an enhancement Expert systems are a program intended to make reasoned to select and test (ST) reasoning algorithm. This algorithm, judgments or give assistance in a complex area in which adjured to be the most precise among the existing diagnostic human skills are fallible or scarce. Considering the work of algorithms, will be enhanced by employing the use of semantic [2] which stated that they are computer system that operates web reasoning structures. Reasoning at the abduction, by applying an inference mechanism to a body of specialist deduction, and induction levels are oriented towards rule base expertise represented in the form of 'knowledge'. They are reasoning pattern in the semantic web. Also, a series of modularized ontology knowledge bases are stacked together in employed as decision support systems, and have many building a complex but distributed knowledge base for the implementation approaches which includes; rule-base entire system. The implementation of this enhanced algorithm (MYCIN and PROSPECTOR), data-base approach, will be used as a test-bed for diagnosing and monitoring descriptive method (INTERNIST and CADUCEUS), and related ailments. Causal Network method. Also, [3] developed the expert system that carries out its diagnoses by organizing symptoms Keywords-semantic web; inference making; ontology; rule into three groups namely Key group(Kg), Sub group(Sg) and set; monitoring; intelligent systems and diagnosis. Unexpected(Ue). Again, [4] designed ASTHMA, an expert system for the diagnoses of asthma. They combined some machine learning algorithms such as Context sensitive auto- I. INTRODUCTION associative memory neural network model (CSAMM), Intelligent medical diagnosis is an area of research that Backpropogation model, C4.5 algorithm, Bayesian Network, leverages on artificial intelligence, and intelligent systems Particle Swarm Optimization to realize their design. Ex-Dr targeted at diagnosis of other systems are becoming Verdis is an integrated expert system [5] Heart Disease pervasive. These intelligent medical systems are also referred Program (HDP) is a medical expert system that enables to as medical expert systems (MES) and the driving force of physicians to enter patient’s symptoms, laboratory tests, and these expert systems are reasoning algorithms. Some of such physical examination. It then generates clinical data that reasoning algorithms are fashioned after mathematical, support the diagnoses of heart disease [6]. statistical, fuzzy and rule-based models. To achieve the Meanwhile, a review of medical reasoning algorithms is reasoning of this enhanced algorithm, the semantic web discussed here. Scheme inductive reasoning algorithm works structures are employed. The semantic web is the power of based on forward thinking [7]. Pattern recognition is inference making and reasoning over ontologically modeled employed in machine learning for assigning some outputs to knowledgebase. Therefore, we use semantic web rule some inputs base on the coordination of a given algorithm languages to encode our rule systems for effective [8]. Hypothetico-deductive reasoning involves generating interoperability with the knowledgebase. As a result, and testing hypotheses in association with the patient’s taxonomies, metadata, classifications, context and ontology presenting symptoms and signs [9] Forward chaining system, have been the basic building blocks of the Semantic Web [1]. includes writing rules to manage sub goals. Whereas, A combination of a formal ontology for the medical domain backward chaining systems automatically manage sub goals with a fine-grained contextual inference making and [10]. Parsimonious Covering Theory (PCT) works on the 5 International Conference on Information and Communication Technology and Its Applications (ICTA 2016) basis of associating a disorder to a set of manifestations [11]. we can conclude that it is correct. The abduction modules Certainty Factor (CF) model is used for managing gets all diagnoses related to symptoms found, and reasons by uncertainty cases in a rule based system [12]. Bayesian hypothesis, studying facts and devising theory to explain it. reasoning algorithms helps in dealing with uncertainties [13]. The process of abduction: The whole process of abduction Fuzzy logic uses linguistic variables to represent operating includes generation, criticism and acceptance of explanatory parameters in order to apply a more human-like way of hypotheses. thinking [14]. Others are: Processing Model for Diagnostic Reasoning [15], Information Processing Approach [16], B. Deduction Select and Test algorithm adjured to be the most approximate [17]. Figure 1 is skeletal model of ST model. Deductive Reasoning is a process in which general premises are used to obtain a specific inference. A form of logic that identifies a particular item by its resemblance to a set of accepted facts. Deductive reasoning moves from a general principle to a specific conclusion. It is inference by reasoning from generals to particulars. Deductions support their conclusions with TRUE result. They compute their results using heuristics. We modify the existing deduction module to be a rule-base deductive reasoning task. Hence, a coordinated rule system and a reasoned are added to semantically realize the deductive reasoning. C. Abstraction The process of mapping descriptive terms understood by patient onto a well-defined symptom entities modeled in the knowledgebase is known as abstraction. In this proposal, we seek to provide patients with a textbox for inputting their entering descriptive terms of how they feel. Our natural language NL-query to Semantic Web SW-query model, then semantically matches their inputs against an ontology of vocabularies in the knowledgebase. The modified abstraction Figure 1. ST Model [17] module allows input to be in speech or textual. Patients may voice in their symptoms and this data will be processed by the voice processor. III. THE PROPOSE ENHANCED SELECT AND TEST (ST) ALGORITHM D. Induction In this section, we present and anatomize our enhanced ST model. The modified model consists of the Abstraction It entails reasoning from the particular to the general. This module and the three logical inference modules namely may or may not be true. But it provides a useful Abduction, Deduction and Induction. More so, the existing generalization. At the induction state, we check if likely ST algorithm data is a-temporal (unable to monitor and store diagnosis meets diagnostic criteria. While Abduction & relevant events, that could support diagnostic reasoning, with Deduction are termed clinical reasoning, Induction is termed respect to time of their occurrence), hence, this made us to clinical decision making. The induction stage in this add a monitoring module to the ST model so as to make it modified ST model builds on the existing features of the data-gathering procedure consistent and as well temporal. existing ST model. Except that we develop a mathematical Contrary to the ST model by [17], we modeled our data model for computing the criticalThreshold parameters, using ontological approach, and the concept of semantic web which is now added to calibrate and alert patient on the rule language is employed for implementing our rule status of the ailment. systems. The following subsections give a breakdown of each of E. Monitoring Agent the consisting components in figure 2. The monitoring agent works continuously in the system. A. Abduction It is more like a daemon which logs events into a Spatial- Temporal-Thematic (STT) ontological database. The essence Our abduction stage improves on the existing abduction of agent is to be able to monitor development of the ailment module. Except that we enabled a semantic web based in the patient’s body, and then adequately signal the needed reasoning operation in the module. We propose the use of alert or logs necessary information that the diagnostic rule engine for this reasoning task. Both the abduction and algorithm will mine data from it. Temporal information the deduction stages here uses this rule engines. And gathered is a clinical data that helps in tracking the compose a rule system for aiding diagnostic reasoning task. progression of a disease in a patient with respect to time. A new parameter, acceptanceThreshold is added to the Spatial information models data that relates with patient and existing likelihoodThreshood parameter. This to check if its environment. Thematic data models concepts and terms every deduction task passes a given acceptance value before used in clinical operations. 6 International Conference on Information and Communication Technology and Its Applications (ICTA 2016) Figure 2. Enhanced Select and Test Model Figure 3 shows the model that the monitoring module works on. On the left hand side, there are four components: the event monitoring, event selector, data gathering, and data IV. THE KNOWLEDGE REPRESENTATION STACK OF THE modeling in spatial-temporal-thematic (STT) format. The ENHANCED ST ALGORITHM event monitor receives information from the intelligent The proposed enhanced Select and Test (ST) algorithm personal agent and then sends it to the event selector which discussed above is a rule base expert system. Figure 4 is a appropriately sieve out the right information to store. Then structural representation of the facts and coordinated rules the data gathering and reasoning faculty generates the system. The structure consists of four layers, and each layer requisite data base on the event. The last component then comprises of facts (model with ontological language OWL) models the data in a STT format. and rules (modeled with semantic web based rule languages, On the right hand side of Figure 3 is the knowledge SWRL). Appended to these four strata is the monitoring modeling in OWL2. This will be fully discussed in chapter agent knowledge model. The first layer models the four. Now, the algorithm for this model will be discussed knowledge of the abstraction layer. Basically, there are three below. modularized data representation. These are a thesaurus Algorithm 1 lists out the steps required for performing modeled with OWL, patient’s personal profile modeled with the monitoring task of the monitoring module. Line 2of the XML, and a rule for mapping descriptive terms of patients algorithm rightly points out the fact that this monitoring into a well-defined symptoms entities, modeled with SWRL. module does its tasks as long as the application is running. The second layer is a knowledge representation for the The module sources its event data by raising some important abduction phase. Knowledge representation at this phase questions at random, and then from the intelligent personal comprises of the facts, modeled with OWL, and rule set for agent which shall be discussed in the next section. Once carrying out abduction (modeled with semantic web rule these data are gathered as indicated by lines 4-5, lines6-12 language SWRL). The deduction module is the next phase cleans the data, formats it into the required style and then for knowledge modeling representation. This phase has a models it in the STT pattern. Afterward, the patient or user is rule set modeled in Jess rule language (JessRL), and the facts alerted in the case of any information that must be passed modeled in OWL also. across. 7 International Conference on Information and Communication Technology and Its Applications (ICTA 2016) Figure 3. Monitoring agent model Figure 4. Knowledge Representation Stack for the Enhanced ST Algorithm 8 International Conference on Information and Communication Technology and Its Applications (ICTA 2016) Algorithm 1: Monitoring module algorithm the responsibility of the induction layer to reason out the correct diagnosis from this three Di. This is achieved by reasoning out which one meet the diagnostic criteria. TABLE I. LIST OF SYMPTOMS Symptom ID Symptoms S1 actinic keratosis: red, pink, or rough patch of skin on sun-exposed areas S2 basal cell carcinoma: raised, waxy, pink bumps S3 squamous cell carcinoma: red, scaly, rough skin lesions, typically on sun-exposed areas such as the hands, head, neck, lips, and ears S4 melanoma: asymmetrically shaped moles or lesions with irregular borders, or change in color or diameter S5 sunburn-like rash that spreads across the nose and both cheeks S6 scaly red patches or ring shapes S7 disc-shaped rash that doesn’t itch or hurt S8 fatigue, headaches, fever, and swollen or painful joints S9 fever, cough, and runny nose S10 reddish-brown rash spreads down the body three to Lastly, the induction phase also comprises of an five days after first symptoms appear ontological knowledge base and a rule set for induction, with S11 tiny red spots with blue-white centers inside the the rule modeled with the Jena rules. The last component in mouth this structured knowledge model is spatial-temporal-thematic S12 papules: small red, raised bumps caused by infected ontological representation of the data generated during the hair follicles monitoring process. S13 pustules: small, red pimples that have pus at their tips S14 nodules: solid, painful lumps beneath the surface of the skin S15 cysts: painful, pus-filled infections found beneath the V. RESULTS AND DISCUSSION skin S16 usually appear on the neck or face of infants As proof of concept, we listed some common symptoms S17 small, red scratch or bump that eventually begins to associated with skin disorders. These symptoms are protrude identifiable with some specific skin disorder disease or S18 most disappear from the skin by age 10 ailment. The proposed enhanced ST algorithm proposed in S19 red, fluid-filled blisters that appear near the mouth this paper help patient to diagnose the particular skin S20 your lips will often tingle or burn before the sore is disorder he/she is suffering from. We note that all Di and Sj visible S21 the sore is painful or tender to the touch are modeled ontologically alongside the necessary rule sets able to help reason out the set of Sj that could cause a particular Di For example, given disease Di, we will need to cyclically reason through abstraction, abduction, deduction, until a TABLE II. LIST OF RELATED SKIN DISORDERS/AILMENT refined result is reasoned out, then we move on to the induction where the real ailment is inductively picked out of Ailment ID Ailment Related Symptoms the few left after the cyclic refinement. D1 Skin Cancer S1, S2, S3, and S4 D2 Lupus S5, S6, S7, and S8 We need to established Di manifesting a set of S, D3 Rubeola (Measles) S9, S10 and S11 once our reasoning structure is able to D4 Acne S12, S13, S14, and S15 accurately establish this set of S, then considering D5 Hemangioma of Skin S16, S17, and S18 D6 Cold Sore S19, S20, and S21 users/patient’s input; a correct diagnoses process is completed. At the Abstraction layer, users input are collected and stored as symptoms. Furthermore, at the abduction layer, VI. CONCLUSION associated skin diseases or disorders of the collected symptoms are reasoned out of the modularized In this paper, we have introduced an enhanced Select and knowledgebase of the abduction layer, alongside its rule sets. Test algorithm which may be employed in diagnosing These skin disorders or disease retrieved at the abduction disease or ailments that have related symptoms. Though it is layer is then sent to the deduction layer. The deduction layer a research in progress, however, this paper has introduced must then intelligent map out the Di manifesting a set of S, the model of the proposed enhanced algorithm. It shows four levels of reasoning: Abstraction, Abduction, Deduction and for every disorder or disease. This cyclic Induction. Furthermore, the knowledge stack of the modeled pattern is continued until it comes to the induction layer. Say was as presented. And finally, a proof of concept was shown at the induction layer, three Di are sent into this module, it is so as to explain the implementation of the algorithm. 9 International Conference on Information and Communication Technology and Its Applications (ICTA 2016) REFERENCES [9] S. P. Kumar, A. Kumar, and V. Sisodia. Clinical Reasoning and Sports Medicine-Application of Hypothetico- Deductive Model”. J [1] R. Cuel, A. Delteil, V. Louis and C. Rizzi. Knowledge Web Sports Med Doping Stud ISSN: 2161-0673 JSMDS, Volume 3 Issue Technology Roadmap. “The Technology Roadmap of the Semantic 1, 2013. Web”, 2008. [10] T. Sharma, N. Tiwari and D. Kelkar. “Study of Difference Between [2] B. S. Todd. “An Introduction to Expert Systems”. Technical Forward and Backward Reasoning”. International Journal of Monograph PRG·95 ISBN 0-902928-73-2, Oxford University Emerging Technology and Advanced Engineering Volume 2, Issue Computing Laboratory Programming Research Group 11 Keble Road 10, October 2012. Oxford OXI3QD England. Pp. 3. 1992. [11] J. Wainer, and A. M. Rezender. A Temporal Extension to the [3] P.S. K. Patra, D. P. Sahu, and I. Mandal. “An Expert System for Parsimonious Covering Theory, 1996. Diagnosis of Human Diseases”. International Journal of Computer Applications, Volume 1 – No. 13, pp. 71-73. 2010. [12] D. E. Heckerman, and E. H. Shortliffe. “From Certainty Factors to Belief Networks”. Artificial Intelligence in Medicine, 1992. [4] B. Prasadl, K.P.E Prasad, and Y. Sagar. “An Approach to Develop [13] J. Gadewadikar, O. Kuljaca, K. Agyepong, E. Sarigul, Y. Zheng, Expert Systems in Medical Diagnosis using Machine Learning and P. Zhang. “Exploring Bayesian Networks for Medical Decision Algorithms (ASTHMA) and a Performance Study”. International Journal on Soft Computing (IJSC), Vol.2, No.1 pp. 26-33. 2011. Support in Breast Cancer Detection”. African Journal of Mathematics and Computer Science Research Vol. 3(10), pp. 225-231, October [5] A. Keles. “Expert Doctor Verdis: Integrated medical expert system”, 2010. Turkish Journal of Electrical Engineering & Computer Sciences, pp. 1032-104, 2014. [14] A. E. Torshabi, M. Riboldi, A. Pella, A. Negarestani, M. Rahnema, and G. Baroni. A Clinical Application of Fuzzy Logic. Fuzzy Logic [6] A. Al-Ajlan. Medical Expert Systems HDP and PUFF. King Saud – Emerging Technologies and Applications, pp. 1-18, 2013. University College of Computer & Information Sciences, Department of Computer Science, pp. 3, 2007. [15] J. Stausberg, and M. Person, “A process model of diagnostic reasoning in medicine”. International Journal of Medical Informatics, [7] K. J. Anderson. Factors Affecting the Development of Undergraduate vol. 54, pp. 9-23, 1999. Medical Students’ Clinical Reasoning Ability. Medicine Learning and Teaching Unit, Faculty of Health Sciences, University of Adelaide, [16] P. M. Wortman. “Medical Diagnosis: An Information-Processing pp. 18, 2006. Approach”. Computers and Biomedical Research, vol. 5, pp. 315-328, 1972. [8] U. A. Umoh, A. A. Umoh, G. G. James, U. U. Oton, and J. J. Udoudo. “Design of Pattern Recognition System for the Diagnosis of [17] I. Fernando, and F. A. Henskens. ST Algorithm for Medical Diagnostic Reasoning. Polibits pp. 23-29, 2013. Gonorrhea Disease”. International Journal of Scientific & Technology Research (IJSTR) Volume 1, Issue 5, pp. 74-79, 2012. 10