Abstract—Sleep Apnea (SA) is one of the most common and important part of sleep disorders. Unfortunately, sleep apnea may be going undiagnosed for years, because of the person’s unawareness. The common diagnose procedure usually required an overnight sleep test. During the test, a recording of many biosignals, which related to breath, are obtained by polysomnography machine to detect this syndrome. The manual process for detecting the sleep Apnea by analysis the recording data is highly cost and time consuming. So, several works tried to develop systems that achieve this automatically. This paper proposes a genetic fuzzy approach for detecting Apnea/Hypopnea events by using Air flow, thoracic and abdominal respiratory movement signals and Oxygen desaturation as the inputs. Results show efficiently of this approach. Index Terms—Sleep disorders, genetic fuzzy algorithm, fuzzy sets. I. INTRODUCTION Study of sleep disorders is important because they are common in general population. For an example, a survey in 1987 [1] reported that at least one symptom of disturbed sleep was present in 41% of all subjects; and still sleep disorder is common now [2], also it was reported by Young that one daytime sleepiness in 5 adults in 2004 [3]. Sleep disorders have several short term and long term side effects[4]. Short term effect leads to impaired attention and concentration, reduce quality of life, increased rates of absenteeism with reduced productivity, and increased the possibility of accidents at work, home or on the road. Long term consequences of sleep deprivation include increased morbidity and mortality from increasing automobile accidents, coronary artery disease, heart failure, high blood pressure, obesity, type 2 diabetes mellitus, stroke and memory impairment as well as depression. Sleep Apnea (SA) is one of the most common and important type of sleep disorders, but because of person’s unawareness, sleep apnea may go undiagnosed for years [5, 6]. Usually SA case is often observed via patient’s spouse or a roommate or a family member who has witnessed the apnea periods alternating with arousals and accompanied by loud snoring [7, 8]. The patients - whom have symptoms of SA – should be examined through an overnight sleep study in sleep center. The diagnosing of SA is usually achieved through using a polysomnography, an integrated device comprising EEG, EMG, EOG, ECG, oxygen saturation [9] airflow Manuscript received July 15, 2012; revised August 31, 2012. The authors are with the University of Technology, Sydney Faculty of Engineering and IT Sydney, Australia (Yashar.Maali@student.uts.edu.au; Adel@eng.uts.edu.au). through the mouth and nose, thoracic and abdominal respiration measurement units [10] and the position of the body during sleep. From overnight sleep studies a respiratory disturbance index (RDI) and an Apnea-Hypopnea index (AHI) will be found, the AHI will holds the summation of apneas and hypopneas per hour, while the RDI will holds the summation of apneas, hypopneas and respiratory arousals values. AHI is used to diagnose and grade the severity of the sleep apnea, according to the Chicago criteria: AHI<5, normal; AHI =5–15, mild; AHI=15–30, moderate; and AHI >30, severe [11]. The polysomnography recording need to be reviewed in order to detect events of SA, the manual review by experts is in fact so cost and time consuming, so several efforts has been done to develop systems that achieve this automatically [12-14]. For this reason several artificial intelligent algorithms are used in this area, like as fuzzy logic that applied to this problem [13, 15, 16]. This paper presents using of fuzzy inference system, but fuzzy rules are generated by genetic algorithms. Using this approach helps to reach a better accuracy and save time that required establishing the fuzzy rules by interview the specialist, especially when they have different opinions about a same case. The rest of this paper is organized as, pre-processing phase and signal analysis and quantification in section 2, basic of proposed fuzzy inference and genetic algorithm reviewed in section 3 and finally, results and conclusion are presented in section 4 and 5 respectively. II. APPROACH AND METHODS In this work, events like “fall in airflow” and “desaturation” or “resaturation” in airflow and oxygen saturation, respectively, are desired. Which these events can be defined as follow: A. Fall in the Air Flow Fall or reduction in Airflow (and also thoracic and abdominal respiratory movement signals) defined as follows: • The mean value of the first minute of the signal is considered as initial value of the normal mean. • Mean of each specified window of signal (for example each 5 seconds) computed, and temporary classified according with the criteria in Table I. • If consecutive windows with duration more than 10s have normal label, the mean value of them will be taken as the new normal mean value. • Else if, consecutive windows with duration more than 10s represents reduce label, it will be considered as fall in signal and restore start of the fall and percentage of it. TABLE I: CORRESPONDING LABELS TO EACH LEVEL OF SIGNAL FALLS Genetic Fuzzy Approach based Sleep Apnea/Hypopnea Detection Yashar Maali and Adel Al-Jumaily International Journal of Machine Learning and Computing, Vol. 2, No. 5, October 2012 685 10.7763/IJMLC.2012.V2.215