TY - GEN
T1 - P-order normal cloud model
T2 - 19th International Conference on Neural Information Processing, ICONIP 2012
AU - Liu, Yu
AU - Zhang, Tianwei
PY - 2012
Y1 - 2012
N2 - Gaussian and power law distribution are two important distribution types. Gaussian distribution is widely spread in nature phenomenon. Most things which obey power law distribution are man-made. We studied the relationship between Gaussian and power law distribution, and gave a new distribution model, is called P-order normal cloud model, of which we proved the basic statistic characteristic. We count distribution law of P-order normal cloud model by experimentation. Then we found a)when P=1, the model obeys Gaussian distribution; 2)with P increased, the characteristics of Gaussian distribution will gradually disappeared, then samples are sharply closed to mean value, but few samples which are distant from mean value are evenly distributed. 3) The samples distribution pattern of high-order normal cloud model(when P>5) reflects power law characteristics.
AB - Gaussian and power law distribution are two important distribution types. Gaussian distribution is widely spread in nature phenomenon. Most things which obey power law distribution are man-made. We studied the relationship between Gaussian and power law distribution, and gave a new distribution model, is called P-order normal cloud model, of which we proved the basic statistic characteristic. We count distribution law of P-order normal cloud model by experimentation. Then we found a)when P=1, the model obeys Gaussian distribution; 2)with P increased, the characteristics of Gaussian distribution will gradually disappeared, then samples are sharply closed to mean value, but few samples which are distant from mean value are evenly distributed. 3) The samples distribution pattern of high-order normal cloud model(when P>5) reflects power law characteristics.
KW - Gaussian distribution
KW - KS statistic
KW - P-order normal cloud model
KW - cloud model
KW - power law
UR - https://www.scopus.com/pages/publications/84869011431
U2 - 10.1007/978-3-642-34481-7_57
DO - 10.1007/978-3-642-34481-7_57
M3 - 会议稿件
AN - SCOPUS:84869011431
SN - 9783642344800
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 467
EP - 474
BT - Neural Information Processing - 19th International Conference, ICONIP 2012, Proceedings
Y2 - 12 November 2012 through 15 November 2012
ER -