Dynamic guard zone for self-supervised learning
نویسندگان
چکیده
Ab.macl The dimension of tbe guard lOne along with its bounds for the Generalised Learning Algorithm (P"thak and Pal. 1986) is determined 1'01' optimum learning. The dimension is found to be dynamic dependlllg on the input sequence and tbe current estimates of classification parameters. Incorporation of tbis higher-order knowledge in a supervisory program improves the system performance. The perfonnance is again found 10 be affecled if the guard zone is shrunk/expanded for 'very weak'l'not 100 weak' estimates when speech d;lta is considered to be mpul
منابع مشابه
Generalized guard-zone algorithm (GGA) for learning: automatic selection of threshold
Absrract-This work IS a continuation of our earlier work on the Generalized Guard-zones Algorithm (GGA) for self-supervised parameter learning. An atlempt is made here for the automatic determination of the guard-zone parameter i' N (i.e. the threshold used for discarding doubtful or mislabeled samples) at every instant of learning, for the general m-elass N-featu re pattern recognition problem...
متن کاملOn the convergence of "A self-supervised vowel recognition system"
AbslraclA self-supervised learning algorithm based on the concept of guard zoncs was developed by Pal et al'l) for studying the adaptive abilHY of a recognition system, starting with non-appropriate representative vectors. Guard zones were used 10 discard unreliable (dou btful) samples from the parameter-updatiug programme, so that the convergenee does not get affccted. The algorithm was implem...
متن کاملEffects of Self-healing on Self-compassion, self-steam and Aggression of Poorly Supervised Teen Girls
Introduction: The aim of this study was to investigate the effectiveness of self-healing training on self-compassion, self-esteem, and aggression in poorly supervised teen girls. Method: The research method was quasi-experimental with a pretest-posttest-follow-up design with a control group. The research population included poorly supervised teen girls in Isfahan that were for sake of this res...
متن کاملA generalised learning algorithm based on guard zones
An algorithm for learning class parameters using a restricted updating programme is described along with investiga tion of its convergence for optimum learning. The algorithm is a generallsation of some existing ones which were found to be useful for practical data. 1. Introduction A similar algorithm was presented by Chien (1970) as a solution to the problem of identifying An adaptive pattern...
متن کاملIschemia detection with a self-organizing map supplemented by supervised learning
The problem of maximizing the performance of the detection of ischemia episodes is a difficult pattern classification problem. The motivation for developing the supervising network self-organizing map (sNet-SOM) model is to exploit this fact for designing computationally effective solutions both for the particular ischemic detection problem and for other applications that share similar characte...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید
ثبت ناماگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید
ورودعنوان ژورنال:
- Pattern Recognition Letters
دوره 7 شماره
صفحات -
تاریخ انتشار 1988