A time series is a set of data recorded over time. For example, we can refer to the time series of the price of a share in the stock market, the amount of rain in an area, etc. One of the most important goals of time series analysis is to predict its future values. Several statistical methods for predicting time series, such as the method of using time series analysis models of Autoregressive integrated moving average model and Seasonal autoregressive integrated moving average model, wavelet analysis methods have been in-troduced. Along with statistical methods, neural networks are also a powerful tool for predicting time series, due to the ability of neural networks to model relationships and complex patterns in data, predicting time series using it has attracted the attention of researchers in various fields and is a research topic. It has become popular. The use of convolutional neural networks to predict time series is known as an e?ective method for analyzing and predicting repetitive patterns in time data. These networks are designed to recognize di?erent patterns in temporal data and can make accurate predictions for future data values using these patterns. Long Short Term Memory (LSTM) Recurrent Neural Networks and Gate Recurrent Neural Networks (GRU) are two types of neural networks used for time series forecasting. LSTM and GRU networks are useful for time series forecasting due to their ability to retain memory over time. These networks, using a memory unit, keep the previous information and make the next prediction according to this information. In this thesis, time series forecasting with convolutional neural network and LSTM and GRU neural networks is investigated. The performance of these networks in prediction accuracy is compared. Also, their performance is compared with statistical methods such as SARIMA.Key ords. Forecasting time series, Neural networks, Recurrent neural networks, Long short-term memory, Gated recurrent unit, Convolutional neural networks.

   Abstract The problem of variable selection is a crucial and difficult aspect of statisticai modeling. Choosing the wrong predictor variables for the final model can result in either underfitting or overfitting. It is crucial to control the rate of false discovery during the inference stage, in variable selection methods. A false discovery rate is said in proportion to auxiliary variables that are improperly chosen among all selected variables. False discovery rate control has been studied for a long time, and several approaches have been developed to achieve an optimal solution. Recently, a new family of FDR control methods called “Knockoff filters” has been introduced. This thesis addresses the issue of variable selection in regression models error false discovery rate control whit various knockoff versions, including fixed-X knockoffs and X-model knockoffs. These knockoff methods can be suitable structures to imitate the existing variables and control the false discovery rate to acceptable limits. The performance of each type of knockoff method will be reviewed in comparison whit each other and some common methods to control the false discovery rate. The efficiency of the presented models has been studied using the data obtained from the simulation.       Keywords: variable selection, sparse regression, knockoff filter, false discovery rate.   

   Diabetes is a   endocrine disorder characterized by chronic hyperglycemia as a result of insulin resistance or deficiency. Diabetes is one of the most common matebolic diseases in the world and one the challenging problems of the present centary, which is the result of interactions between genetic and behavioral predisposition and environmental factors. Considering the prevalence of type 2 diabetes around the world, it is useful for doctors to identify connections and discore new rules, and for this reason, doctors and researchers analyzed and investigated the cause of the growing increase of this disease by using artificial intelligence and its subset. In this thesis, by using machine learning, including decision tree, random forest, logistic regression and neural network, we analyze and investigate the diagnosis and prognosis of type 2 diabetes by using the data if Ravansar cohort, which includes 10047 people and 137 variables, we understand that the main factors affecting this disease are age, fasting blood sugar level, manganese, selenium, etc. Also, people who are involved with this disease   should change their lifestyle according to the doctor's advice so that they don't face more problems in the rest of their lives. Keywords: Type 2 diabetes, Artificial Intelligenece, Machine Learning, Decision Tree, Random Forest, Logistic Regression, Neural Network.

   In this paper, the aim is to investigate the dependent competing risk model under middle censoring using some bivariate distributions such as Marshall-olkin bivariable Weibull distribution and Marshall-olkin bivariable Gompertz distribution and to compare the maximum likelihood estimation approach and Bayesian approach. To achieve this, subset of real data as well as simulations using R software are used, and the maximum likelihood estimates and Bayesian estimation are compared with each other.

The intuition of risk is based on two main concepts: the possibility of a negative outcome, i.e. a lo   and the variability in terms of an expected result, i.e. a deviation. Since the time when the modern theory of finance was accepted, the role of risk measurement has attracted attention. Initially, it was predominantly used as a dispersion measure, such as variance, which contemplates the second pillar of the intuition. More recently, the occurrence of critical events has turned the attention to tail-risk measurement, as is the case of well known Value at Risk (VaR) and Expected Shortfall (ES)measures, which contemplate the first pillar. In this Thesis, a risk measure is considered which contemplate both pillars of intuition on risk. These pillars include the possibility of negative results and variability over an expected result, as a single measure. This resulting composition, based on properties of the two components, is a coherent risk measure. Similar results for the cases of convex and co-monotone risk measures are exposed. Then, the eleven well-known risk measures consider from different classes. Finally, the empirical values of corresponding to loss, deviation and loss-deviation risk measures are obtained and compared using Monte Carlo and real data.   

  The subject of classification is the one of the important issues indifferent sciences. The logistic regression is the one of the statistical methods to classify data in which the underlying distribution of the data is assumed to be known. Today, researchers in addition to statistical methods use other methods such as machine learning to classify data.  In this thesis, the decision trees C4.5, C5, CART, CHAID, and QUEST  are introduced,  and each of them is completely studied. Some ensemble learning algorithms such as random forest, Bagging, and Boosting in the field of supervised learning are also explained. Finally,  using five data sets, we compare the performance of these algorithms with respect to the accuracy measure.

Gaussian time-series models are often specified through their spectral density. Such models pose several computational challenges, in particular because of the non-sparse nature of the covariance matrix. We derive a fast approximation of the likelihood for such models. We use importance sampling to correct for the approximation error. We show that the variance of the importance sampling weights vanishes as the sample size goes to infinity. We show that the posterior is typically multi-modal, and derive a Sequential Monte Carlo sampler based on an annealing sequence in order to sample from the approximate posterior. Performance of the overall approach is evaluated on simulated and real datasets.

   Normal 0 false false false EN-US X-NONE FA The weighted k out of n system is a one with n components that each component has a specific weight and it works if sum of its active components be at least k. A lot of researches and studies have been carried out around the properties of k out of n systems. This thesis investigates the reliability and some properties of these systems. One of these properties is the redundancy allocation that its result has been presented in the first part of this thesis. In the next part, the result of a weighted k out of n system with dependent components have been studied. Finally, k out of n system generalized to the weighted (k1, k2, …, km) out of n system and its result have been presented.

  در آمار يكي از مهم ترين ابزارها براي تجزيه و تحليل داده ها ، به دست آوردن برآورد مناسب يك تابع است كه روش هاي مختلفي براي آن به وجود آمده است. يكي از معروف ترين و ساده ترين روش هاي برآورد، روش كمترين مربعات معمولي است كه در شرايط مطلوب مزيت هاي فراواني دارد. اين روش در رگرسيون بعد بالا كاربردي ندارد و دليل اين مساله هم اين است كه در رگرسيون بعد بالا به علت زياد بودن متغير هاي پيشگو، سبب دشوار شدن تفسير مدل و كاهش دقت در برآورد مي شود. در چنين شرايطي محقق مي تواند با كاهش متغير هاي پيشگو و درواقع حذف متغير هاي كم اثر با استفاده از روش هاي خاصي كه بيان مي گردند، سبب بهتر شدن تفسير اين مدل ها شود. دراين پايان نامه ابتدا ضمن معرفي روش هايي كه به روش هاي انقباضي معروف هستند، روش هايي همانند لاسو ، لاسو گروهي، ريج، بريج و الاستيك نت مورد بررسي قرار مي گيرند. از طرف ديگر   مدل هاي با ضرايب متغير نيز از جمله مهم ترين ابزارها براي كشف الگوهاي حركتي در بسياري از علوم از جمله: سرمايه گذاري مالي، اپيدميولوژي، علوم سياسي،علوم پزشكي، اكولوژي و غيره هستند. اين مدل ها بسط طبيعي مدل هاي كلاسيك پارامتري هستند كه با تفسير پذيري خوب، محبوبيت زيادي در تجزيه و تحليل داده ها به دست آورده اند. انعطاف پذيري و تفسير پذيري بالاي اين مدل ها در دهه اخيرسبب شده كه تحولات شگرف و جالبي در روش شناسي، نظري و كاربردي در اين زمينه پديد آيد. در ادامه ضمن معرفي مدل هاي با ضرايب متغير به بررسي مهم ترين روش برآورد پارامتر در اين مدل ها كه روش استفاده از تابع هسته است پرداخته مي شود. همچنين مدل هاي با ضرايب متغير تعميم يافته را معرفي خواهيم كرد و با استفاده از روش رگرسيون بريج در مدل هاي بعد بالا، برآورد پارامتر ها در اين حالت را نيز مورد بحث و بررسي قرار خواهيم داد. در نهايت انتخاب متغير براي مدل هاي با ضريب متغير تعميم يافته بعد بالا را مورد بررسي قرار مي دهيم كه از روش انقباضي لاسو گروهي در اين بحث استفاده مي كنيم ، همچنين براي انتخاب بهترين زير مجموعه از متغير هاي موجود از روش اطلاع بيزي تعميم يافته استفاده مي كنيم و تمام اين مسائل را در نرم افزار R مورد بحث و بررسي قرار خواهيم داد.

Waiting lines and service efficiency are the important elements for any bank. Queuing theory has been fairly a successful tool in the performance analysis of waiting lines. In this paper, an optimized model is proposed to improve the bank queuing system based on queuing theory. This method can optimize the number of server and improve the service efficiency that could effectively cut down service costs and customer’s waiting time  

In many environmental applications involving spacial and spatio-temporal data, covariance functions are the fundamental tools for modeling dependent data which observed over space and (or) time. Constructing nonseparable spatio-temporal variogram and covariance functions is one of the issues related to spatio-temporal geostatistics for space-time data. Limitations on the number of locations and high values of parameters motivate the need for practical and efficient method to estimate the parameters of covariance models for spatial interpolation, or kriging. In this paper, we discuss different types of spatio-temporal data, also we combine these methods with Bees algorithm (BA) to explore the spatiotemporal correlation structure of monthly ground-water data in Ilam city. Finally, we predict the level of some spacial stations using spatio-temporal kriging which the efficiency of model was assessed by Mean of Kriging Standard Deviations (MKSD). Keywords: Kriging, Spatio-temporal data, Bee Algorithm

  در اين رساله، انواعي از سياست نگهداري براي پيشگيري خسارت­هاي(گاهاً جبران­ناپذير) ناشي از خرابي قطعات در حين كار كردن، مطالعه مي­شود. همانطور كه بيان شد شكست واحدها در زمان كار كردن ممكن است گران تمام شود. در مواردي كه با افزايش عمر، نرخ شكست افزايش مي­يابد، براي جلوگيري از خسارات، واحد را بايد قبل از آنكه خراب شود تعويض كنيم. سياست تعويض عمر، از پايه­هاي اصلي سياست نگهداري است كه براي جلوگيري از شكست يك واحد در زمان كار كردن است. در سياست تعويض عمر، قطعه در زمان شكست و يا زمان مشخص t تعويض مي شود اگر قطعه در زمان t در حال كار كردن باشد.مشخص است كه تعيين زمان t يكي از مسايل بسيار مهم در بحث سياست جايگذاري است.بارلو و پروشان(1965) ميانگين زمان شكست تحت سياست جايگذاري را به عنوان معياري از تاثير t معرفي نمودند.

In many practical studies in medical, social and economic studies, the data are collected over time as discrete longitudinal data. One of the important issues in this data is the effect of the previous response on the current response, which can be considered by a transition model. Another important issue in the data set is heterogeneity among individuals which can be covered by adding a random effect to the transition model. Another important problem in longitudinal data is missing values. In this thesis, a non-ignorable missingness mechanism is considered for modeling binary and ordinal longitudinal data by a random effects transition model. Some simulation studies are performed for illustration of the proposed approaches; also, the proposed approaches are used for analyzing a real data set.

برآورد تابع بقا و عوامل موثر بر بيماران مبتلا به سرطان پستان در عراق از سال 2014 تا 2016  

  اخيراً‏، مسئله‌ي نقطه‌ي تغيير يكي از موضوعاتي است كه مورد توجه بسياري از آماردانان واقع شده است. نقطه‌ي تغيير‏، مي‌تواند در بسياري از زمينه‌هاي تحقيقاتي و صنعتي مفيد واقع شود و از خسارات جبران‌ناپذير پيشگيري كند. در اين رساله‏، ما به مطالعه‌ي مسئله‌ي نقطه‌ي تغيير و تشخيص آن‏، با استفاده از آزمون فرض پرداخته‌ايم. ابتدا‏، به تعريف نقطه‌ي تغيير و مثال‌هايي از آن مي‌پردازيم‏، سپس تشخيص تك‌نقطه‌ي تغيير در داده‌هاي طول عمر با حجم كم را مورد بررسي قرار مي‌دهيم. در راستاي اجراي اين آزمون فرض‏، از آماره‌هاي متفاوتي استفاده مي‌كنيم كه يكي از آن‌ها‏، آماره‌ي آزمون نسبت درستنمايي تجربي است. بالاكريشنان و همكاران ‎‎‎‎(2016) ‎‏ تشخيص تك‌نقطه‌ي تغيير و بررسي آن را در دو حالت پارامتري و ناپارامتري مورد مطالعه قرار داده‌اند. هدف اين رساله‏، تشخيص نقطه‌ي تغيير و مقايسه‌ي آماره‌هاي متعدد از طريق توان آزمون‌هاست.كليد واژه: آزمون فرض، نقطه‌ي تغيير، نسبت درستنمايي تجربي، داده‌هاي طول عمر

Recently, one of the subjects that is attractive for statisticians on applied fields, is optimal designs to do the experiments. Optimal designs are obtaind using real-valued function which is call optimal criteria.Statistical models to study the behavior of absorption or Drog release are called statistical models in pharmacokinetic/pharmacodynamics. In this thesis, we consider optimal designs for models in pharmacokinetic/pharmacodynamics studies.Since that information matrixs for these models depend on the unknown parameters, locally optimal designs have been considered to find optimal designs. Numerical results have been obtained for D-, A-, E-optimal designs.

  In this thesis, we give a new distribution of a new class of distributions called the Hyperbolic Cosine – F (HCF) distribution.Since the Weibull distribution of the most popular and most widely used distributions in reliability and lifetime data analysis. A special case of this family is the Hyperbolic Cosine Weibull distribution . We also discuss simulation method, also maximum likelihood minimum spaccings estimators of the parameters of the distribution.And continue   we have discussed   with the second type of censorship increasingly on the distribution, to estimate the parameters. The new distribution can be used effectively in the analysis of survival data.

Dynamic formulations of reliability theory are of so considerable interest in the analysis and comparison of working systems in real time. Dynamic signatures, which have been considered by many researchers in the recent years, are useful tools in the studying and comparisons of the lifetimes of used systems.In this thesis, we study the dynamic signatures and its application in reliability theory of coherent systems. In this direction, we first introduce the signature of a coherent system having i.i.d components lifetimes and investigate some their properties and applications in reliability theory of coherent systems. Then, we studythe dynamic signatures under some information about the lifetime of the system. Samaniego et al. (2009) considered a situation that some partial information about the system and its components lifetimes are available and defined the dynamic signatures under such information.The last purpose of this thesis is introducing of such dynamic signatures, which depends only on the system structure, andstudyingtheir applications.

In many from life ‎testing and reliability ‎studies, the experimenter may not always obtain complete information on failure times for all experimental units. for Example, individuals in a clinical trial may drop out of the study, may have to be terminated for lack of funds or in an industrial experiment, units may break accidentally. ‎Therefore, one has to remove some units prior to failure for saving time and cost associated with testing. Data obtained from such experiments are called censored. The most common censoring schemes are type I and type II. The Type I and Type II censoring schemes have major deficiency in that they only do not allow removal of units at points other than the terminal point of the experiment. Due to experimenter use a versatile scheme of censoring called progressive.‎This thesis has been focused on the scenario of progressive Type II censoring. A problem associated with this scheme is that the total testing duration might be unacceptably long. To address this issue, a hybrid variant of the ‎progressive ‎‎censoring scheme was proposed in which imposing a time limit T on the test. Although this hybrid ‎progressive ‎‎censoring scheme controls the total testing duration not larger than T, it is possible that the effective sample size is very small or even zero in which usual efficient statistical inference may not be feasible. To strike a balance between the ‎total ‎testing ‎time ‎and ‎the ‎efficiency ‎in ‎statistical ‎inference, ‎Ng ‎et.al ‎(2009)‎ proposed an adaptive Type II progressive censoring scheme.In this thesis influential methods for progressive Type II censoring and adaptive Type II progressive censoring under some statistical distributions are studied. We ‏‎obtain‎ both maximum likelihood estimators, approximate maximum likelihood estimators and observed information matrix for the unknown parameters. We simulated values of the estimates parameters, a comparison of the values variance and covariance of the estimators with those obtained from the corresponding observed information matrix and coverage probabilities for pivotal quantities based estimators. Various interval estimation methods for the unknown parameters such as asymptotic confidence intervals with both observed information matrix and fisher information matrix, percentile bootstrap and bootstrap-t confidence intervals are obtained. Then these methods are compared in terms of their expected lengths and coverage probabilities using simulation.

  The allocation of redundant component(s) in a system so as to optimize the lifetime of the system is quite ‌considerable inreliability theory. In engineering systems, the k-out-of-n system is a very important structure, whichworks if and only if at least k components are operating. In this thesis, we consider the problems of optimal allocation of R redundancies to n components of   k-out-of-n systemsand series systems, as particular case of k-out-of-n systems when k=n,with respect to some stochastic orderings.Two commonly used methods to allocate redundancy are active redundancy and standby redundancy. We discuss the allocation of active redundancies to series and k-out-of-n systems in the three cases with i.i.d components and redundancies, i.i.d components and i.i.d redundancies and stochastically ordered components. We also compare the lifetimes of series systems arising out of one standby redundancy. Optimal allocation of redundancies in -out-of-   ystems

  In the recent years, many researches on dynamic measures of uncertainty have been carriedout which indicate their practical importance. One of the related interesting issues is thenonparametric estimation of such measures. Since in the reliability and survival analysis,samples usually contain censored data, so in this thesis, using the kernel density estimation,we investigate the nonparametric estimation of generalized past entropy and Renyi’s residualentropy based on the censored data. Some of the asymptotic properties of the estimators arestudied.