Feature selection is one of the most important techniques in machine learning and pattern recognition, which eliminates redudant features and selects a suitable subset of features. This avoids overfitting when building the model and improves the model performance. In many applications, obtaining labeled data is costly and time consuming, while unlabeled data are readily available. Therefore, semi-supervised feature selection methods can be used to consider both labeled and unlabeled data in the feature selection process. In this paper, a semi-supervised Sparse feature selection method is proposed based on hessian Regularization and Fisher discriminant analysis which selects the appropriate features using the labeled data and the local structure of both labeled and unlabeled data. In the proposed method, an objective function based on semi-supervised scatter matrix and l2,1-norm is presented for feature selection which considers the correlation among features. To solve the proposed objective function, an iterative algorithm is used and its convergence is experimentally and theoretically proved. The results of the experiments on five data sets indicate that the proposed method improves the selection of relevant features compared to other methods used in this paper.

Advanced Sparse coding-based image fusion methods use some prior information to fuse low-resolution multispectral (LR-MS) and panchromatic images to create a high-resolution multispectral image (HR-MS). This information mainly includes a sparsity term, spectral unmixing, and nonlocal similarities. These prior terms are usually considered in the Sparse optimization problem as constraints with specific Regularization parameters. During the optimization, the Regularization parameter of each prior term is optimized by considering the other two prior terms as constants. This study aims to simultaneously optimize the Regularization parameters of prior terms in a Sparse coding image fusion method to construct an HR-MS from input LR-MS and Pan images. Several optimization methods, including particle swarm optimization, ant colony optimization, differential evolution, and genetic algorithm were used to optimize the Regularization parameters. The results showed that particle swarm optimization had the highest performance in increasing the peak signal-to-noise ratio on the dataset available from the study area. The advantages of the proposed optimized Sparse coding (OSC) approach are the ability to, 1) preserve spatial details while eliminating spectral distortions, 2) simultaneously optimize the Regularization parameters of prior terms in a Sparse coding image fusion framework, 3) considering nonlocal similarities to enhance fusion result, and 4) promising fusion results over heterogeneous regions with highly spectral variations. The relative dimensionless global error in synthesis, spectral angle mapper, universal image quality index, and peak signal to noise ratio criteria were at least 0. 76, 1. 16, 0. 0257, and 2. 68 better than those achieved by conventional PS methods, i. e., Gram-Schmidt, Brovey transform, generalized intensity-hue-saturation, smoothing filter-based intensity modulation, and a novel Sparse coding-based image fusion method. According to the results, better preservation of spatial details and lower spectral distortions can be achieved using the proposed OSC approach.

Gravity inversion methods play a fundamental role in subsurface exploration, facilitating the characterization of geological structures and economic deposits. In this study, we conduct a comparative analysis of two widely used Regularization methods, Tikhonov (L2) and Sparse (L1) Regularization, within the framework of gravity inversion. To assess their performance, we constructed two distinct synthetic models by implementing tensor meshes, considering station spacing to discretize the subsurface environment precisely. Both methods have proven ability to recover density distributions while minimizing the inherent non-uniqueness and ill-posed nature of gravity inversion problems. Tikhonov Regularization yields stable results, presenting smooth model parameters even with limited prior information and noisy data. Conversely, Sparse Regularization, utilizing sparsity-promoting penalties, excels in capturing sharp geological features and identifying anomalous regions, such as mineralized zones. Applying these methodologies to real gravity data from the Safu manganese deposit in northwest Iran, we assess their efficacy in recovering the geometry of dense ore deposits. Sparse Regularization demonstrates superior performance, yielding lower misfit values and sharper boundaries during individual inversions. This underscores its capacity to provide a more accurate representation of the depth and edges of anomalous targets in this specific case. However, both methods represent the same top depth of the target in the real case study, but the lower depth and density distribution were not the same in the XZ cross-sections. Inversion results imply the presence of a near-surface deposit characterized by a high-density contrast and linear distribution, attributed to the high grade of manganese mineralization.

One of the challenges of convolutional neural networks (CNNs), as the main tool of deep learning, is the large volume of some relevant models. CNNs, inspired form the brain, have millions of connections. Reducing the volume of these models is done by removing (pruning) the redundant connections of the model. Optimal Brain Damage (OBD) and Sparse Regularization are among the famous methods in this field. In this study, a deep learning model has been trained and the effect of reducing connections with the aforementioned methods on its performance has been investigated. As the proposed approach, by combining the OBD and Regularization methods its redundant connections were pruned. The resulting model is a smaller model, which has less memory and computational load than the original model, and at the same time its performance is not less than the original model. The experimental results show that the hybrid approach can be more efficient than each of the methods, in the most tested datasets. In one dataset , with the proposed method, the number of connections were reduced by 76%, without sacrificing the efficiency of the model. This reduction in model size has decreased the processing time by 66 percent. The smaller the software model, the more likely it is to be used on weaker hardware, found everywhere, and web applications.

One of the most important social problem after world warII in developing country is rapid urbanization. In most of the developing country yearly growth urban population is among 5 untill 8 percent. This urban explosive grow and the slums are effect of the inside immigration from village to urban, that call in various place and various form like marginal, squatter, Illegal, Irregular, spontaneous, unauthorized, informal settlement.Our country (IRAN) like other developing country encounter with this problem. At present in most urban and industrial city like Arak city there are slums.These thesis investigate structural and cultural feature of slums (Bagh Khalaj district) that doing with document and surveying at first abstract of science literature (definitions, scores and feature) of slums in other country and Iran and then survey physical and humanly structure of Arak city.After that survey slums and especially Bagh Khalaj district (case study) in Arak city. At the end of thesis on the base of results and to point out strategy, limitations, facilities and problems and then present solutions.

Ill-conditioned linear systems may frequently arise in discretization of integral equations and many other real world applications. Solving such systems by classical methods might fail or result to solutions that are meaningless from practical point view. Moreover a slight perturbation in the right hand side vector might also lead to an enormous change of the solution vector due to ill-conditioned ness. To find meaningful solutions of such systems, the Tikhonov Regularization is an effective technique that has been widely used. In this paper we use it to solve ill-conditioned linear systems and also to find closest feasible linear systems to nearly feasible linear systems by smallest changes in problem data. Throughout the paper numerical results are reported to demonstrate the practical efficiency of the presented algorithms.