• 충청수학회지
• /
• 제36권2호
• /
• pp.65-76
• /
• 2023

With the Stirling matrix S of the second kind and the Pascal matrix T, we study recurrence rules and sequences of certain sums over the matrix ST^k. We find a matrix E satisfying ST = ES and interrelations of S and the Stirling matrix of the first kind.

• 응용통계연구
• /
• 제24권2호
• /
• pp.373-381
• /
• 2011

본 논문은 실험자료에 대한 분석모형으로 이원 분산분석모형을 가정한다. 고정효과 모형의 가정하에 요인별 변동량을 구하기 위한 방법으로 제1종 분석을 다루고 있다. 모형의 순차적 적합에 따라 얻어지는 요인별 제곱합의 계산방법으로 대수적 방법이 아닌 사영에 의한 분석방법을 제공한다. 관측자료를 다차원상의 공간벡터로 간주할 때, 최소 제곱법에 의한 요인별 변동량은 계획행렬로 생성되는 모수추정 공간에서 요인별 부분공간으로의 사영에 이르는 거리 제곱으로 구해질 수 있음을 논의하고 있다. 또한 사영행렬로 부터의 고유벡터와 고유근을 이용하여 요인별 변동량을 구하는 방법을 제공하고 있다. 균형자료나 불균형자료에서 모형의 순차적 적합에 따른 제1종 분석이 행해질 때 요인별 변동량의 합은 처리제곱합과 일치하나 제2종 분석의 경우 불균형자료에서 이러한 성질이 만족되지 않음을 논의하고 있다.

• 음성과학
• /
• 제10권4호
• /
• pp.189-200
• /
• 2003

The objective of this study was to examine two new parameters by which we could screen people with malignant vocal folds. The new parameters were the difference sums and Pearson correlation coefficients between adjacent pairs of intensity level matrices of narrow-band spectra. Audio files from the Korean Disordered Speech Database were analyzed by Praat, a speech analysis software, to obtain matrices of 400 intensity levels at 16 time points of each sustained vowel spectra. We limited our study to 12 normal subjects and 20 patients with malignant vocal folds who recorded at least three Korean vowels at a sound-proofed booth in Busan National University Hospital. Results indicated that the average coefficients of the abnormal subjects were much lower than those of the normal subjects while the average difference sums of the patients were much higher than those of the normal ones. Also, we found that the degree of the malignancy of the vocal folds was related to the coefficients and sums. However, some subjects at the initial stages of cancerous vocal folds yielded almost comparable coefficients and difference sums to those of the normal speakers. Further studies on larger databases will be desirable to set certain criteria or threshold levels for screening people with vocal fold diseases.

• Journal of the Korean Data and Information Science Society
• /
• 제25권1호
• /
• pp.19-26
• /
• 2014

다항 로지스틱 회귀모형의 설명변수가 연속형과 범주형을 모두 포함할 때 범주형 설명변수는 직합을 적용하고 연속형 설명변수는 텐서곱을 적용하는 모형을 제안한다. 변수선택의 기준으로 BIC를 사용하고, 제안된 모형의 알고리즘을 구현하였다. 구현된 알고리즘을 실제 자료에 적용하여 기존의 방법과 비교하여 제안된 모형이 더 좋은 분류율을 보임을 확인하였다.

• 대한수학회지
• /
• 제40권5호
• /
• pp.757-768
• /
• 2003

In 1967, as an answer to the question of P. Erdos on a set of integers having distinct subset sums, J. Conway and R. Guy constructed an interesting sequence of sets of integers. They conjectured that these sets have distinct subset sums and that they are close to the best possible with respect to the largest element. About 30 years later (in 1996), T. Bohman could prove that sets from the Conway-Guy sequence actually have distinct subset sums. In this paper, we generalize the concept of subset-sum-distinctness to k-SSD, the k-fold version. The classical subset-sum-distinct sets would be 1-SSD in our definition. We prove that similarly derived sequences as the Conway-Guy sequence are k-SSD.

• 호남수학학술지
• /
• 제35권3호
• /
• pp.351-372
• /
• 2013

In this paper, we study a distinction the two generating functions : ${\varphi}^k(q)=\sum_{n=0}^{\infty}r_k(n)q^n$ and ${\varphi}^{*,k}(q)={\varphi}^k(q)-{\varphi}^k(q^2)$ ($k$ = 2, 4, 6, 8, 10, 12, 16), where $r_k(n)$ is the number of representations of $n$ as the sum of $k$ squares. We also obtain some congruences of representation numbers and divisor function.

• 대한수학회지
• /
• 제55권5호
• /
• pp.1207-1220
• /
• 2018

The generalized hyperharmonic numbers $h^{(m)}_n(k)$ are defined by means of the multiple harmonic numbers. We show that the hyperharmonic numbers $h^{(m)}_n(k)$ satisfy certain recurrence relation which allow us to write them in terms of classical harmonic numbers. Moreover, we prove that the Euler-type sums with hyperharmonic numbers: $$S(k,m;p):=\sum\limits_{n=1}^{{\infty}}\frac{h^{(m)}_n(k)}{n^p}(p{\geq}m+1,\;k=1,2,3)$$ can be expressed as a rational linear combination of products of Riemann zeta values and harmonic numbers. This is an extension of the results of Dil [10] and $Mez{\ddot{o}}$ [19]. Some interesting new consequences and illustrative examples are considered.

• 대한수학회지
• /
• 제43권4호
• /
• pp.815-828
• /
• 2006

Let {$V_{nk},\;k\;{\geq}\;1,\;{\geq}\;1$} be an array of rowwise independent random elements which are stochastically dominated by a random variable X with $E\|X\|^{\frac{\alpha}{\gamma}+{\theta}}log^{\rho}(\|X\|)\;<\;{\infty}$ for some ${\rho}\;>\;0,\;{\alpha}\;>\;0,\;{\gamma}\;>\;0,\;{\theta}\;>\;0$ such that ${\theta}+{\alpha}/{\gamma}<2$. Let {$a_{nk},k{\geq}1,n{\geq}1$) be an array of suitable constants. A complete convergence result is obtained for the weighted sums of the form $\sum{^\infty_k_=_1}\;a_{nk}V_{nk}$.

• 대한수학회보
• /
• 제52권1호
• /
• pp.35-43
• /
• 2015

Let q > 1 be an odd integer and c be a fixed integer with (c, q) = 1. For each integer a with $1{\leq}a{\leq}q-1$, it is clear that the exists one and only one b with $0{\leq}b{\leq}q-1$ such that $ab{\equiv}c$ (mod q). Let N(c, q) denote the number of all solutions of the congruence equation $ab{\equiv}c$ (mod q) for $1{\leq}a$, $b{\leq}q-1$ in which a and $\bar{b}$ are of opposite parity, where $\bar{b}$ is defined by the congruence equation $b\bar{b}{\equiv}1$ (modq). The main purpose of this paper is using the mean value theorem of Dirichlet L-functions to study the mean value properties of a summation involving $(N(c,q)-\frac{1}{2}{\phi}(q))$ and Kloosterman sums, and give a sharper asymptotic formula for it.

• International Journal of Contents
• /
• 제10권1호
• /
• pp.18-22
• /
• 2014

Hidden nodes have a key role in the information processing of feed-forward neural networks in which inputs are processed through a series of weighted sums and nonlinear activation functions. In order to understand the role of hidden nodes, we must analyze the effect of the nonlinear activation functions on the weighted sums to hidden nodes. In this paper, we focus on the effect of nonlinear functions in a viewpoint of information theory. Under the assumption that the nonlinear activation function can be approximated piece-wise linearly, we prove that the entropy of weighted sums to hidden nodes decreases after piece-wise linear functions. Therefore, we argue that the nonlinear activation function decreases the uncertainty among hidden nodes. Furthermore, the more the hidden nodes are saturated, the more the entropy of hidden nodes decreases. Based on this result, we can say that, after successful training of feed-forward neural networks, hidden nodes tend not to be in linear regions but to be in saturated regions of activation function with the effect of uncertainty reduction.