1Principle Machine Learning Engineer
Azure Sovereign Operations
Microsoft, Redmond, Washington, USA
Email: [email protected]
Page Number: 1-30
Training supervised machine learning models requires labeled examples. A judicious choice of examples
is helpful when there is a significant cost associated with assigning labels. This article improves upon
a
promising extant method โ Batch-mode Expected Model Change Maximization (B-EMCM) method โ for
selecting examples to be labeled for regression problems. Specifically, it develops and evaluates
alternate
strategies for adaptively selecting batch size in B-EMCM
By determining the cumulative error that occurs from the estimation of the stochastic gradient descent,
a
stop criteria for each iteration of the batch can be specified to ensure that selected candidates are
the most
beneficial to model learning. This new methodology is compared to B-EMCM via mean absolute error and
root mean square error over ten iterations benchmarked against machine learning data sets.
Using multiple data sets and metrics across all methods, one variation of AB-EMCM, the max bound of the
accumulated error (AB-EMCM Max), showed the best results for an adaptive batch approach. It achieved
better root mean squared error (RMSE) and mean absolute error (MAE) than the other adaptive and
non-adaptive batch methods while reaching the result in nearly the same number of iterations as the
non-adaptive
batch methods.
keywords: active learning (AL), batch mode, expected model change, linear
regression, nonlinear
regression
We consider the problem of finding a pair of functions ฮฆ(๐ฅ,๐ก)and ๐ค(๐ฅ,๐ก)that satisfy the
equation
๐ค๐ฅ๐ฅ(๐ฅ,๐ก) + ๐ค๐ก๐ก(๐ฅ,๐ก) = ฮฆ(๐ฅ,๐ก) under Cauchy boundary conditions. We will see that an
approximate
solution can be found using the techniques of generalized inverse problem of moments and find dimensions
for the error of the estimated solution
keywords: generalized moment problem; integral equations; Poisson equation, inverse
source
Graph theory recently considers as a modern field of mathematics and it was introduced by a great
mathematician Leonhard Euler in 1735. Late then it is a flowered in the strong tool used in closely each
area
of science and nowadays it is most attractive and active area of mathematics research. Actually, graph
theory
is the study of graph, structure and molecules which have association in each other. It is very
essential part of
discrete mathematics which has been created with collaboration of nodes (vertices )and edges (lines or
links),
if a graph is connected with vertices directly by edges it is direct graph or a graph having
symmetrically lines
is also called direct graph or diagraph. In the studies of graph, we have investigated so many kinds of
graph,
Hypo-Hamiltonian is one of them. It is most important and it can be defined as the graph which hasnโt
Hamiltonian cycle but can be developed with the removing of single vertex form the developed graph. Lot
of
researchers have given the contribution in this research just like ๐๐๐๐๐ ๐๐ก ๐๐ has worked on โA
Two-Connected Graph with ๐บ๐๐๐๐๐โ๐ Propertyโ containing 12,18 and 25 vertices.In this research
paper we have
developed two different graphs consisting 25 Eulerian graph which has highest cycle and path order
๐ถ(๐บ) =
24 and ๐(๐บ) = 25.Other graph is in 3D under cycle which contains 23 nodes.
keywords: Group theory, longest path and longest Cycle, Gallaiโs
Property
In telepsychiatric treatments, the medical data of various types of patientsโ data are to be
communicated in a well-protected way between the patient and the psychiatrists. Homeopathy drugs
play a significant role in treating such psychiatric patients. The objective of this manuscript is to
develop a strong session key with the help of neural cryptography. Neural cryptography is the
extensive usage of the artificial neural networks for the cryptographic functions. E-medicines were
encrypted here by the classical methods and the proposed session keys of 128 bits long. The length
has been selected to 128 bits in order to have less memory space needed. Alzheimer's disease,
Dementia, Insomnia, Stress, OCD, Brain restlessness, Depression, etc are the most clinical mental
complications observed in this post COVID-19 situations. This is the era of digital health to deal the
remote patients who are non-invasive in nature. In this manuscript, a session key has been proposed
by the Key Distribution Centre (KDC) with the aid of artificial neural networks. Strong session keys
were derived of 128 bits length which in turn used for AES 128 encryption. To avoid the
Man-InThe-Middle attacks, it has been designed that the server machine shall accumulate all the needful
symptoms of the psychiatric patients.
keywords: Psychiatry, Homeopathy E-medicines, Neural Cryptography, Secure
Transmission
In this work a deterministic model of bird flu epidemic was investigated. The model included
susceptible,
infected, and removed or recovered birds. This model revealed that a non-zero population with some peak
and low periods had non-zero constants that made the infectious population to have a possible limit it
cannot exceed, no matter the rate of infection. We observed that the disease-free steady state of our
model
was globally asymptotically stable. We also observed that there should be a bound at which susceptible
become infected.
keywords: Deterministic Model, Bird Flu Epidemic, Infectious Population, Disease Free Steady
State
and Globally Asymptotically Stable.
This paper presents the estimates of older-adult (persons aged 60 years and above) mortality from
census data in Nigeria using the fixed-base model, an application based on the variable-r
method. From the 1991 and 2006 censuses in Nigeria, it was observed that the probability of dying
between ages 60 and 65 [
5 60 q
] is 0.068 (or 68 per 1000 persons) among males and 0.033 (or 33
per 1000 persons) among females. While the probability of dying between ages 60 and 70 [
6010 q
]
is 0.221 (or 221 per 1000 persons) among males and 0.256 (or 256 per 1000 persons) among
females. Furthermore, the probability of dying between ages 60 and 75 [
6015 q
] is 0.324 (or 324
per 1000 persons) among males and 0.389 (or 389 per 1000 persons) among females. The results
appear more like the pattern of mortality rates in developed countries than in developing (subSaharan)
countries, probably due to the lack of completeness of female census data. Even at that,
the results from the model are consistent with the estimates from other known sources. It is
expected that the results will help in tracking the progress of Sustainable Development Goals
(SDGs) in Nigeria, among others
keywords: Older-adult mortality, census, fixed-base model, probability of dying, Nigeria.
The present review article is a enhancement of fixed point theorems for two pairs of compatible type (B)
mappings with two pairs of weakly compatible mappings in Banach space. The present paper divided into
two parts, in first part we have generalized a common fixed point theorem for two pairs of compatible
type (B) mappings and in second part we mapped a common fixed point theorem for two pairs of weakly
compatible mappings in Banach space using a special type contractive condition usually named square
inequality.
keywords: Compatible mapping of type (B), Common fixed point theorem, Weakly
Compatible
mapping, Square inequality, Banach space
We have conducted analytical investigations to study the influence Hall current and Buoyancy
distribution on (MHD) unsteady heat and mass transfer free convective in vertical channels. Appropriate
dimensional quantities were used in changing the dimensional governing partial differential equations to
non-dimensional form. An analytical method of variables separable was employed in finding the
analytical solution of dimensionless governing partial differential equations. The expressions of
velocity,
temperature, concentration, skin friction, Nusselt number as well as Sherwood number were gotten
analytically and discussed using line graph. From the result obtained, it was observed that velocity
profile
enhances with the increase of Grasshop number (Gr), Mass Grasshop number (Gc), Hall current (m) and
Buoyancy parameter (rt
). While it reduces with the increase of Magnetic parameter (M) and Pradlt
number (Pr). Similarly, temperature profile enlarges with the increase of Radiation parameter (R) and
Buoyancy parameter (rt
). While the opposite behaviour was observed with the increase of Pradlt number
(Pr). Concentration profile diminishes with the increase of both Chemical reaction parameter (Kr
) and
Schmidt number (Sc). Skin friction gets enhanced with the increase of and Buoyancy parameter (rt
) at
both y=0 and y=1 and gets reduced with the increase of Hall current parameter (m) at both y=0 and y=1.
Nusselt number enlarges with the increase of Radiation parameter (R) at both y=0 and y=1. While it gets
lowered with the increase of Buoyancy parameter (rt
) at both y=0 and y=1. Sherwood number slightly
increase with the increase of chemical reaction parameter (Kr
) at both y=0 and y=1 and significantly
decrease with the increase of Schmidt number (Sc) at both y=0 and y=1.
Babyemi, A.W1, Onwuka G.I2, Okeleke S.U3, and B.
Shehu4
1,2Department of Mathematics and Statistics
Kebbi State University of Science and Technology,
Aliero
3,4Department of Mathematics and Statistics,
Kebbi State University of Science and Technology, P.M. B. 1144, Aliero
Corresponding Author Email: [email protected]
Page Number:111-122
Scholars over decades have argued and debated over the volatility behaviour of assets returns, but
unilateral conclusion is yet to be reached as regard some unique features of some market assets. More
so, the fact that prices of asset fluctuate more frequently over a period of time than the ones in the
usual
markets, as such, the present study investigated the asset return volatility in Nigerian Stock Exchange
Market using five different firms which these include GUINESS, UBA, UBN, CADBURY and FIRST
BANK. The data used for this study were daily from 4th January, 2010 to 16th December, 2019 and
sourced from the Nigerian Stock Exchange. The research employed Generalized Auto Regressive
Conditional Heteroskedasticity (GARCH) model, and other developed members of its family-type, such
as Autoregressive Fractionally Integrated Moving Average (ARFIMA), Fractionally Integrated
Generalized Auto Regressive Conditionally Heteroskedastic (FIGARCH) models. These competing
GARCH-type family models were tested and selection of the optimum model was carried out using the
Log likelihood (LogL), Akaike information criterion (AIC), Schertz information criterion (SIC) and
Hannan Quinn criterion (HQC). The results from the analysis revealed that FIGARCH (1,d,1) produced
the best fit for GUINESS, FBANK, UBN and CADBURY, where the fraction order are (0.34, 0.28,
0.32 and 0.24) respectively. While ARFIMA (1,1,1) produced the best fit for UBA. This clearly
indicated that asymmetric FIGARCH model produced better fits in volatility models.
keywords: Garch, Figarch, Arfima, Arch Lm, Volatility
The study aimed at investigating the effects of demographic parameters on both economic and
population growth in Nigeria. Three models were employed in the study and results from model 1
depicted that economic growth has a positive effects to BR and negatively affected by DR, these
results had demonstrated that increase in population growth in Nigeria is favorable to economic
growth of the nation while death was found unfavorable to the economic growth in Nigeria. This
result is an indication of the fact that Nigeria is not facing the problem of overpopulation; rather the
capacities of Nigerian Government and responsible organizations to create a favorable economic
environment by channeling the right resources into the right place. In the second model we also
discovered that labor force was statistically significant with a P-value of 0.00328. Thus, model 3
regressed GDP, labor force, health expenditure and corruption perception on population growth. The
results depicted that all, except health expenditure with 0.82552 P-value, are statistically
significant.
The results have shown the influence of economic growth, labour force and corruption on the growth
of population in Nigeria. The data in this work were of two types from three different sources;
National Bureau of Statistics (NBS), African Development Bank (ADB) and World Bank (WB), first
part ranges from 1995 to 2018 and second part ranges from 1985 to 2018
keywords: ARDL, Unit Root, GDP, Death Rate (DR), Birth Rate (BR), Labor Force (LP),
Health
Expenditure (HE), Net Migration (NM), Corruption Perception (CP), and Population Growths (PG).
