Monaural source separation is a challenging issue due to the fact that there is only a single channel available; however, there is an unlimited range of possible solutions. In this paper, a monaural source separation model based hybrid deep learning model, which consists of convolution neural network (CNN), dense neural network (DNN) and recurrent neural network (RNN), will be presented. A trial and error method will be used to optimize the number of layers in the proposed model. Moreover, the effects of the learning rate, optimization algorithms, and the number of epochs on the separation performance will be explored. Our model was evaluated using the MIR-1K dataset for singing voice separation. Moreover, the proposed approach achieves (4.81) dB GNSDR gain, (7.28) dB GSIR gain, and (3.39) dB GSAR gain in comparison to current approaches
Acquisition provisions in Islamic jurisprudence
The construction of embankment for roadway interchange system at urban area is restricted due to the large geometry requirements, since the value of land required for such construction is high, and the area available is limited as compared to rural area. One of the optimum solutions to such problem is the earth reinforcement technique which requires a limited area for embankment construction. Gypseous soil from Al-Anbar governorate area was obtained and subjected to various physical and chemical analysis to determine it is properties. A laboratory model box of 50x50x25 cm was used as a representative embankment; soil has been compacted in five layers at maximum dry density (modified compaction) and an aluminum reinforcement strips we
... Show MoreThe internet, unlike other traditional means of communication, has a flexibility to stimulate the user and allows him to develop it. Perhaps, the reason for the superiority of the internet over other traditional means of communication is the possibility of change and transmission from one stage to another in a short period. This means that the internet is able to move from the use to the development of the use and then the development of means and innovation as the innovation of the internet is a logical product of the interaction of the user with the network. The internet invests all the proposals and ideas and does not ignore any even if it is simple. This is represented in social networking sites which in fact reflects personal emotio
... Show MoreAbstract:
In this research we discussed the parameter estimation and variable selection in Tobit quantile regression model in present of multicollinearity problem. We used elastic net technique as an important technique for dealing with both multicollinearity and variable selection. Depending on the data we proposed Bayesian Tobit hierarchical model with four level prior distributions . We assumed both tuning parameter are random variable and estimated them with the other unknown parameter in the model .Simulation study was used for explain the efficiency of the proposed method and then we compared our approach with (Alhamzwi 2014 & standard QR) .The result illustrated that our approach
... Show MoreChemotherapy is one of the most efficient methods for treating cancer patients. Chemotherapy aims to eliminate cancer cells as thoroughly as possible. Delivering medications to patients’ bodies through various methods, either oral or intravenous is part of the chemotherapy process. Different cell-kill hypotheses take into account the interactions of the expansion of the tumor volume, external drugs, and the rate of their eradication. For the control of drug usage and tumor volume, a model based smooth super-twisting control (MBSSTC) is proposed in this paper. Firstly, three nonlinear cell-kill mathematical models are considered in this work, including the log-kill, Norton-Simon, and hypotheses subject to parametric uncertainties and exo
... Show MoreThe aim of the research is to examine the multiple intelligence test item selection based on Howard Gardner's MI model using the Generalized Partial Estimation Form, generalized intelligence. The researcher adopted the scale of multiple intelligences by Kardner, it consists of (102) items with eight sub-scales. The sample consisted of (550) students from Baghdad universities, Technology University, al-Mustansiriyah university, and Iraqi University for the academic year (2019/2020). It was verified assumptions theory response to a single (one-dimensional, local autonomy, the curve of individual characteristics, speed factor and application), and analysis of the data according to specimen partial appreciation of the generalized, and limits
... Show MoreIn this work, two groups of nanocomposite material, was prepared from unsaturated polyester resin (UPE), they were prepared by hand lay-up method. The first group was consisting of (UPE) reinforced with individually (ZrO2) nanoparticles with particle size (47.23nm). The second group consists of (UPE) reinforced with hybrid nanoparticles consisting of zirconium oxide and yttrium oxide (70% ZrO2 + 30% Y2O3) with particles size (83.98nm). This study includes the effect of selected volume fraction (0.5%, 1%, 1.5%, 2%, 2.5%, 3%) for both reinforcement nano materials. Experimental investigation was carried out by analyzing the thermo-physical properties like thermal conductivity, thermal diffusivity and specific heat for the polymeric composit
... Show MoreAbstract: The increased interest in developing new photonic devices that can support high data rates, high sensitivity and fast processing capabilities for all optical communications, motivates a pre stage pulse compressor research. The pre-stage research was based on cascading single mode fiber and polarization maintaining fiber to get pulse compression with compression factor of 1.105. The demand for obtaining more précised photonic devices; this work experimentally studied the behavior of Polarization maintaining fiber PMF that is sandwiched between two cascaded singe mode fiber SMF and fiber Bragg gratings FBG. Therefore; the introduced interferometer performed hybrid interference of both Mach-Zehnder
... Show More