IDENTIFICA TION OF THE MATURITY LEVEL OF MANGO "ARUMANIS" USING ARTIFICIAL NEURAL NElWORK Atris Suyantohadi·, Guntarti Tatik Mulyati a, Wahyu Supartono· Titik F.Djafar b a Department of Agricultural Induslrial Technology. Gadjah Mada University Bulaksumur. Yogyakarta.Indonesia b Agricultural Research and Development Institutions, Yogyakarta, Indonesia Abstract: The sensory perception technique to detennine the criterion of maturity level of mango 'arumanis' has not been able to predict fruit quality convincingly. This research aims to decide maturity levels and quality of the mangoes using artificial neural network teclutiquc. The non-destructive analysis on fruitC<)lor, lcngtll and diameter from the selected mango population at various ftuitsses froro ' tbll beginning to the ripe mangoes shows that there are changes in the analy@d salllplesi Tlteehanges also show by lheblln of fruit textureftfidchcmjcal 'US3rand Rcid, p" s .. ta .. . J. C. h.ThO of . Vtl .. . .• ·. Jn . s. .• . .• network modeicclJ.n the· appHotdon i'fO,DlUloan ldentuy lileotlterinof lbo UUll18oe8; UtlriPCfJ:lOt 1ll8tureonouab, maturity and ripe.Copy right I gOOl IFAC. Keywo:tdll : IdentlncaUon. artificial ncuml network, maturity level 1. INTRODUCTION Mosl fonnel1 in Indoneilft; the maturity level of Ill8JJIOes wlt41M IOl1IOty percepUon. This method take, muchthnoMd tbe . f.CfWt is not reliable (Panlastisco, 1992).111e ·· fllaturitylcwcJ has been influenced by the physical and chemical factors (Yuniarti, 2000; Satuhu, 1997). This research aims to de':.ide the maturity levels and quality of mango 'arumanis' non--destructively using the artificial neural network. The method to detennine the maturity of mangoes done by the sensory perception tcchnique will be calculated quantitatively with the measurement tool of laboratory. The physical parameter of fruit with non..(iestructive test like fruit color, length, thickness and density are the input factors. The output factor is the fruit quality based on the maturity level criterion. Using the method, the complexity of the relationship between the output input factors that reflects the maturity level pattern can be simulated and monitored based on the criterion of the fruit. The detennination of the maturity level and quality uses the design of artificial 325 ncuml network model roptnces the seftSo!), perception technique. 2. METHOD AND MA'fERIALS This research relates with the filed rescarcll and laboratory research WiOl mango 'arumanis' as the object. The field research is to analysiS the development level of the mango from the yOWlg (unripe) mango till it ripens itself. The laboratory research is to analyse some factors that are capable to change during the fruit developlflent and to examine the related reference. The data from both researchcs are for the knowledge source and the learning of neural network design in identifying the maturity level and the quality of mango 'arumnnis' 2.1. Research Equipment and Materulls The primary material of the research is mango 'arumanis' variety taken from the mango plantation in Watugajall, Yogy akarta , Indonesia . The