A three-stage learning algorithm for deep multilayer perceptron (DMLP) with effective weight initialisation based on sparse auto-encoder is proposed in this paper, which aims to overcome difficulties in training deep neural networks with limited training data in high-dimensional feature space. At the first stage, unsupervised learning is adopted using sparse auto-encoder to obtain the initial weights of the feature extraction layers of the DMLP. At the second stage, error back-propagation is used to train the DMLP by fixing the weights obtained at the first stage for its feature extraction layers. At the third stage, all the weights of the DMLP obtained at the second stage are refined by error back-propagation. Network structures and values of learning parameters are determined through cross-validation, and test datasets unseen in the cross-validation are used to evaluate the performance of the DMLP trained using the three-stage learning algorithm. Experimental results show that the proposed method is effective in combating overfitting in training deep neural networks.
Objective(s): assessment of the health follow up and weight control for women with osteoporosis and find out the relationship between their health follow up and weight control and their socio-demographic characteristics.
Methodology: A descriptive study was conducted on women with osteoporosis for the period of September, 26th 2020 to Jun, 20th 2021. Non- probability (convenient) sample of (70) women with osteoporosis selected from (5) Private Clinics for Joints and Fractures in Baqubah City. A questionnaire was designed though extensive review of literatures and it consists of three parts: the first part includes women’s socio demographic characteristics, the second part inclu
... Show MoreAmong the metaheuristic algorithms, population-based algorithms are an explorative search algorithm superior to the local search algorithm in terms of exploring the search space to find globally optimal solutions. However, the primary downside of such algorithms is their low exploitative capability, which prevents the expansion of the search space neighborhood for more optimal solutions. The firefly algorithm (FA) is a population-based algorithm that has been widely used in clustering problems. However, FA is limited in terms of its premature convergence when no neighborhood search strategies are employed to improve the quality of clustering solutions in the neighborhood region and exploring the global regions in the search space. On the
... Show MoreAbstract
Although the rapid development in reverse engineering techniques, 3D laser scanners can be considered the modern technology used to digitize the 3D objects, but some troubles may be associate this process due to the environmental noises and limitation of the used scanners. So, in the present paper a data pre-processing algorithm has been proposed to obtain the necessary geometric features and mathematical representation of scanned object from its point cloud which obtained using 3D laser scanner (Matter and Form) through isolating the noised points. The proposed algorithm based on continuous calculations of chord angle between each adjacent pair of points in point cloud. A MATLAB program has been built t
... Show MoreThe successful implementation of deep learning nets opens up possibilities for various applications in viticulture, including disease detection, plant health monitoring, and grapevine variety identification. With the progressive advancements in the domain of deep learning, further advancements and refinements in the models and datasets can be expected, potentially leading to even more accurate and efficient classification systems for grapevine leaves and beyond. Overall, this research provides valuable insights into the potential of deep learning for agricultural applications and paves the way for future studies in this domain. This work employs a convolutional neural network (CNN)-based architecture to perform grapevine leaf image classifi
... Show MoreIn this article, the high accuracy and effectiveness of forecasting global gold prices are verified using a hybrid machine learning algorithm incorporating an Adaptive Neuro-Fuzzy Inference System (ANFIS) model with Particle Swarm Optimization (PSO) and Gray Wolf Optimizer (GWO). The hybrid approach had successes that enabled it to be a good strategy for practical use. The ARIMA-ANFIS hybrid methodology was used to forecast global gold prices. The ARIMA model is implemented on real data, and then its nonlinear residuals are predicted by ANFIS, ANFIS-PSO, and ANFIS-GWO. The results indicate that hybrid models improve the accuracy of single ARIMA and ANFIS models in forecasting. Finally, a comparison was made between the hybrid foreca
... Show MoreBackground: This study aims to investigate the effect of fixed orthodontic appliances and/or antihypertensive drugs on the weight of experimental rats. Materials and Methods: Thir-ty-six male Wistar albino rats were subjected to a split-mouth design study, in which an orthodontic appliance was inserted in one side to move the first molar mesially for 2 weeks while the other side acted as a control to tooth movement. The rats were allocated into three groups: group A (n = 12), without any pharmacological treatment; group B (n = 12), subcu-taneous injection of bisoprolol fumarate (5 mg/kg) daily; and group C (n = 12), subcutaneous injection of valsartan (10 mg/kg) daily. A fixed orthodontic appliance with a closing coil spring delivering 5
... Show MoreIn this paper, the botnet detection problem is defined as a feature selection problem and the genetic algorithm (GA) is used to search for the best significant combination of features from the entire search space of set of features. Furthermore, the Decision Tree (DT) classifier is used as an objective function to direct the ability of the proposed GA to locate the combination of features that can correctly classify the activities into normal traffics and botnet attacks. Two datasets namely the UNSW-NB15 and the Canadian Institute for Cybersecurity Intrusion Detection System 2017 (CICIDS2017), are used as evaluation datasets. The results reveal that the proposed DT-aware GA can effectively find the relevant features from
... Show MoreMost of the recent works related to the construction industry in Iraq are focused on investigating the validity of local raw materials as alternatives to the imported materials necessary for some practical applications, especially in thermal and sound insulation. This investigation includes the use of limestone dust as partial substitution of cement in combination with foam agent and silica fume to produce sustainable Lightweight Foam Concrete (LWFC). This study consists of two stages. In the first stage, trial mixes were performed to find the optimum dosage of foam agent. Limestone dust was used as a partial replacement for cement. Chemical analysis and fineness showed great similarity with cement. Many concrete mixes were prepared
... Show MoreThis research aims at building a proposed training program according to the self-regulated strategies for the mathematics teachers and to identify the effect of this program on relational Mathematics of teachers. The sample of the research was (60) Math teachers; (30) teachers as experimental group and (30) teachers as control group. The results of the current research reacheded that the proposed training program according to some self-managed learning strategies, meets the needs of trainees with remarkable effectiveness to improve the level of their teaching performance to achieve the desired goals. Training teacher according to self-managed learning strategies is effective in bringing about the transition of training to their students
... Show More