دانلود و نمایش مقالات مرتبط با Prediction::صفحه 9
دانلود بهترین مقالات isi همراه با ترجمه فارسی 2

با سلام خدمت کاربران در صورتی که با خطای سیستم پرداخت بانکی مواجه شدید از طریق کارت به کارت (6037997535328901 بانک ملی ناصر خنجری ) مقاله خود را دریافت کنید (تا مشکل رفع گردد). 

نتیجه جستجو - Prediction

تعداد مقالات یافته شده: 932
ردیف عنوان نوع
81 پیش بینی قیمت بیت کوین با استفاده از یادگیری ماشین: یک رویکر برای مهندسی ابعاد نمونه
سال انتشار: 2020 - تعداد صفحات فایل pdf انگلیسی: 13 - تعداد صفحات فایل doc فارسی: 32
پس از فراز و فرودهای قیمت های ارزهای رمزنگاری شده در سال های اخیر، بیت کوین به صورت فزاینده ای به عنوان یک دارایی برای سرمایه گذاری در نظر گرفته شده است. به خاطر ماهیت بسیار بی ثبات قیمت بیت کوین، لازم است تا پیش بینی های مناسبی صورت گیرد تا، بر اساس آن، بتوان در مورد سرمایه گذاری تصمیم گیری نمود. با وجودی که تحقیقات جاری برای پیش بینی دقیق تر قیمت بیت کوین از یادگیری ماشین استفاده کرده اند، تعداد اندکی از آنها به امکان استفاده از تکنیک های مختلف مدل سازی برای نمونه هایی با ساختار داده ای و ویژگی های بعدی مختلف توجه کرده اند. به منظور پیش بینی بهای بیت کوین در فرکانس های مختلف با استفاده از تکنیک های یادگیری ماشین، ابتدا قیمت بیت کوین را بر اساس قیمت روزانه و قیمت فرکانس بالا طبقه بندی می کنیم. مجموعه ای از ویژگی های با ابعاد بالا از جمله دارایی و شبکه، معاملات و بازار، توجه و قیمت لحظه ای طلا برای پیش بینی قیمت روزانه بیت کوین استفاده می شود، در حالی که ویژگی های اصلی تجارت که از تبادل ارز رمزنگاری شده حاصل شده اند، برای پیش بینی قیمت در فواصل 5 دقیقه ای استفاده می شوند. روشهای آماری شامل رگرسیون لجستیک و آنالیز افتراقی خطی برای پیش بینی قیمت روزانه بیت کوین با ویژگی های ابعاد بالا، به دقت 66٪ رسیده و از الگوریتم های یادگیری پیچیده تر ماشین پیشی می گیرند. در مقایسه با نتایج مبنا برای پیش بینی قیمت روزانه، با بالاترین دقت در روش های آماری و الگوریتم های یادگیری ماشینی، به ترتیب 66٪ و 3/65٪، به عملکرد بهتری دست پیدا می کنیم. مدلهای یادگیری ماشینی، شامل جنگل تصادفی ،XGBoost، آنالیز افتراقی درجه دو، ماشین بردار پشتیبان و حافظه کوتاه مدت بلند برای پیش بینی قیمت 5 دقیقه ای بیت کوین که دقت آنها به 67.2% رسیده است، از روشهای آماری بهتر هستند. بررسی ما در مورد پیش بینی قیمت بیت کوین را می توان مطالعه ای مقدماتی در مورد اهمیت ابعاد نمونه در تکنیک های یادگیری ماشین در نظر گرفت.
کلمات کلیدی: مهندسی ابعاد نمونه | اصل Occam’s Razor | پیش بینی قیمت بیت کوین | الگوریتم های یادگیری ماشین
مقاله ترجمه شده
82 Comparison of the impacts of empirical power-law dispersion schemes on simulations of pollutant dispersion during different atmospheric conditions
مقایسه تأثیر برنامه های پراکندگی قانون تجربی قدرت در شبیه سازی پراکندگی آلاینده در شرایط جوی مختلف-2020
Accurate and rapid predictions of air pollutant dispersion are important for effective emergency responses after sudden air pollution accidents (SAPA). Notably, dispersion parameters (σ) are the key variables that influence the simulation accuracy of dispersion models. Empirical dispersion schemes based on power-law formulas are probably appropriate choices for simulations in SAPA because of the requirement for only routine meteorological data. However, performance comparisons of different schemes are lacking. In this study, the performances during simulations of air pollutant dispersion of four typical empirical parameterised schemes, i.e. BRIGGS, SMITH, Pasquill-Gifford, and Chinese National Standard (CNS), were investigated based on the GAUSSIAN plume model with datasets for the classic Prairie Grass experiments, 1956. The performances when simulating peak and overall concentrations in different Pasquill atmospheric stability classes (A, B, C, D, E, F) were quantitatively analysed through different statistical approaches. Results showed that the performances of four schemes for peak and overall concentrations were basically consistent. Scheme CNS in unstable atmospheric conditions (A, B, and C) performed significantly better than the others according to performance criteria, which included the lowest mean of absolute value of fractional biases, lowest normalised mean square errors, and largest mean values of the fraction within a factor of two when predicting peak and overall concentrations, respectively. Schemes BRIGGS and P-G exhibited slightly better performances during the neutral condition (D) followed by scheme CNS. Schemes SMITH and CNS demonstrated slight merits in predicting concentrations compared to the other schemes during stable conditions (E and F). As a whole, scheme CNS generally performed well for the different atmospheric stability classes. These analysis results can help to fill in the data gaps and improve our understanding of the influence of typical power-law function schemes on simulations of air pollutant dispersion. The results are expected to provide scientific support for air pollution predictions, especially during emergency responses to SAPA.
Keywords: Empirical power-law dispersion schemes | Atmospheric stability | Performance evaluation | Statistical analysis | Emergency response | Sudden air pollution accidents
مقاله انگلیسی
83 Laminar flame speeds of methane/air mixtures at engine conditions: Performance of different kinetic models and power-law correlations
سرعت شعله چند لایه مخلوط های متان / هوا در شرایط موتور: عملکرد مدل های مختلف جنبشی و همبستگی قدرت قانون-2020
The laminar flame speed is an important input in turbulent premixed combustion modelling of spark ignition engines. At engine-relevant temperatures and pressures, its measurement is challenging or not possible and thereby it is usually obtained from simulations based on chemical models or power-law correlations. This work aims to investigate the performance of different models and power-law correla- tions in terms of predicting laminar flame speeds of methane/air at engine conditions. The propagation of spherically expanding laminar flames in a closed chamber was simulated and laminar flame speeds were computed over a broad range of pressures (1-120 atm) and temperatures (30 0-110 0 K) for methane/air mixtures based on seven kinetic models. It was found that at engine conditions, there are notable dis- crepancies among the predictions. GRI Mech. 3.0 and USC Mech. II respectively predict the largest and smallest values at high pressure conditions. This was explained by the difference in CH 3 oxidation and recombination according to reaction pathway analysis. Additionally, laminar flame speeds of methane flames were experimentally determined under engine-relevant conditions. It was shown that the recently developed Foundational Fuel Chemistry Model Version 1.0 model predicts closely the data at high pres- sures and temperatures. Therefore, it was chosen as the reference model for the comparisons. Thirteen published power-law correlations for laminar flame speeds of CH 4 /air were implemented, and their per- formance in predicting the laminar flame speeds at engine conditions was investigated. Most of these correlations have been derived for a narrow range of temperatures and pressures, which are lower than those encountered in engines. A new power-law correlation was derived based on predictions by the Foundational Fuel Chemistry Model Version 1.0. This new correlation is expected to provide reliable pre- dictions at engine conditions for a stoichiometric methane/air mixture and thereby it is recommended to be used in modeling turbulent premixed combustion in spark-ignition engine simulations.
Keywords: Laminar flame speed | engine conditions | methane | power-law correlation | propagating spherical flame
مقاله انگلیسی
84 Using reinforcement learning for maximizing residential self-consumption - Results from a field test
استفاده از یادگیری تقویتی برای به حداکثر رساندن خود مصرفی مسکونی - نتایج یک آزمون میدانی-2020
This paper presents the results from a real residential field test in which one of the objectives was to maximize the instantaneous self-consumption of the local photovoltaic production. The field test was part of the REnnovates project and was conducted in different phases on houses in several residential districts located in Soesterberg, Heerhugowaard, Woerden and Soest, the Netherlands. To maximize self- consumption, buffered heat pump installations for domestic hot water and stationary residential battery systems were chosen due to their respective thermal and electrical storage capacities. The algorithm used to tackle the associated sequential decision-making problem was model-based reinforcement learning. The proposed algorithm learns the stochastic occupant behavior, uses predictions of local photovoltaic production and considers the dynamics of the system. The results show that this algorithm increased the average self-consumption percentage of the local PV generation (used instantaneously in situ ) on average by 14%, even if only buffered heat pump installations for domestic hot water were used. This increase was achieved without causing any perceived discomfort to the residential end users. The average energy shifted per day from the solar production period to the night by the 2 kW/3.6 kWh batteries was 1.5 kWh. The main contribution of this work was therefore the real field implementation of the proposed algorithm. The results demonstrate that it is possible to improve even further the integration of local production using flexible loads.
Keywords: Reinforcement learning | Q-Learning | Field test | Solar PV generation | Thermal storage | Thermostatically controlled loads | Electrical storage | Battery | Residential loads
مقاله انگلیسی
85 Study on deep reinforcement learning techniques for building energy consumption forecasting
مطالعه تکنیک های یادگیری تقویتی عمیق برای پیش بینی مصرف انرژی در ساخت-2020
Reliable and accurate building energy consumption prediction is becoming increasingly pivotal in build- ing energy management. Currently, data-driven approach has shown promising performances and gained lots of research attention due to its efficiency and flexibility. As a combination of reinforcement learning and deep learning, deep reinforcement learning (DRL) techniques are expected to solve nonlinear and complex issues. However, very little is known about DRL techniques in forecasting building energy con- sumption. Therefore, this paper presents a case study of an office building using three commonly-used DRL techniques to forecast building energy consumption, namely Asynchronous Advantage Actor-Critic (A3C), Deep Deterministic Policy Gradient (DDPG) and Recurrent Deterministic Policy Gradient (RDPG). The objective is to investigate the potential of DRL techniques in building energy consumption predic- tion field. A comprehensive comparison between DRL models and common supervised models is also provided. The results demonstrate that the proposed DDPG and RDPG models have obvious advantages in forecast- ing building energy consumption compared to common supervised models, while accounting for more computation time for model training. Their prediction performances measured by mean absolute error (MAE) can be improved by 16%-24% for single-step ahead prediction, and 19%-32% for multi-step ahead prediction. The results also indicate that A3C performs poor prediction accuracy and shows much slower convergence speed than DDPG and RDPG. However, A3C is still the most efficient technique among these three DRL methods. The findings are enlightening and the proposed DRL methodologies can be positively extended to other prediction problems, e.g., wind speed prediction and electricity load prediction.
Keywords: Energy consumption prediction | Ground source heat pump | Deep reinforcement learning | Asynchronous advantage Actor-Critic | Deep deterministic Policy gradient | Recurrent deterministic Policy gradient
مقاله انگلیسی
86 Imbalanced credit risk evaluation based on multiple sampling, multiple kernel fuzzy self-organizing map and local accuracy ensemble
ارزیابی ریسک اعتباری نامتوازن بر اساس نمونه گیری چندگانه ، نقشه خود سازماندهی فازی چند هسته ای و گروه دقت محلی-2020
Credit risk evaluation model is generally regarded as a valid method for business risk management. Although the most of literatures about credit risk evaluation always use class-balanced data as sample sets, the study on class-imbalanced datasets is more suitable for actual situation. This paper proposes a new ensemble model to evaluate class-imbalanced credit risk, which integrates multiple sampling, multiple kernel fuzzy self-organizing map and local accuracy ensemble. To preprocess imbalanced sample sets of credit risk evaluation, multiple sampling approaches (synthetic minority over-sampling technique, under sampling and hybrid sampling) are improved and integrated to acquire balanced datasets. To construct more suitable base classifiers, multiple kernel functions (Gaussian, Polynomial and Sigmoid) respectively are used to improve fuzzy self-organizing map. Then, the balanced sample sets are respectively processed by the improved base classifiers to acquire different prediction results. The local accuracy ensemble method is employed to dynamically synthesize these prediction results to obtain final result. The new ensemble model can further avoid over-fitting and information loss, be more suitable to handle the dataset including different financial indicators, and acquire the stable and satisfactory prediction result for imbalanced credit risk evaluation In the empirical research, this paper adopts the financial data from Chinese listed companies, and makes the comparative analysis with the relative models step by step. The results can prove that the new ensemble model presented by this article has better performance than other methods in terms of evaluating the imbalanced credit risk.© 2020 Elsevier B.V. All rights reserved.
Keywords: Credit risk evaluation | Class-imbalanced data | Multiple sampling | Multiple kernel fuzzy self-organizing map | Local accuracy ensemble
مقاله انگلیسی
87 Effect of moisture content on thermal decomposition and autoignition of wood under power-law thermal radiation
تأثیر رطوبت بر تجزیه حرارتی و اتوماسیون چوب تحت قانون تابش حرارتی -2020
Pyrolysis and autoignition of beech wood, with moisture content (MC) from 0% to 38%, exposed to power-law thermal radiation is investigated. An experimental facility capable of irradiating time-varying heat flux was utilized to conduct bench-scale tests. An analytical model was proposed to estimate the ignition time, surface and in-depth temperatures. Meanwhile, a transient numerical solver, FireFOAM, was employed to simulate the experimental measurements. The results show that surface and in-depth temperatures rise more quickly with larger heat flux and lower MC. Although the analytical model overestimates surface temperature slightly due to the ignoring of pyrolysis, both the analytical and numerical predictions are at acceptable levels. In-depth temperatures cannot be accurately predicted due to the recondensation of the migrated water vapor beneath the evaporation layer. No substantial discrepancy is found between the 25% and 38% MC surface temperatures since significant cracks emerge on surface upon heating. The average measured ignition temperature is 395 °C which is used as the ignition criterion. Ignition time increases with lower heat flux and larger MC when MC < 25%, but no further increase is observed for MC > 25% due to the crack effect. The predicted ignition times match the experimental results relatively well despite some minor deviations.
Keywords: Thermal decomposition | Auto-ignition | Beech wood | Moisture content | Time-varying heat flux | FireFOAM
مقاله انگلیسی
88 Detecting personal microbiota signatures at artificial crime scenes
تشخیص امضاهای میکروبیوت شخصی در صحنه های جرم ساختگی -2020
When mapped to the environments we interact with on a daily basis, the 36 million microbial cells per hour that humans emit leave a trail of evidence that can be leveraged for forensic analysis. We employed 16S rRNA amplicon sequencing to map unique microbial sequence variants between human skin and building surfaces in three experimental conditions: over time during controlled and uncontrolled incidental interactions with a door handle, and during multiple mock burglaries in ten real residences. We demonstrate that humans (n = 30) leave behind microbial signatures that can be used to track interaction with various surfaces within a building, but the likelihood of accurately detecting the specific burglar for a given home was between 20–25%. Also, the human microbiome contains rare microbial taxa that can be combined to create a unique microbial profile, which when compared to 600 other individuals can improve our ability to link an individual ‘burglar’ to a residence. In total, 5512 discriminating, nonsingleton unique exact sequence variants (uESVs) were identified as unique to an individual, with a minimum of 1 and a maximum of 568, suggesting some people maintain a greater degree of unique taxa compared to our population of 600. Approximate 60–77% of the unique exact sequence variants originated from the hands of participants, and these microbial discriminators spanned 36 phyla but were dominated by the Proteobacteria (34%). A fitted regression generated to determine whether an intruder’s uESVs found on door handles in an office decayed over time in the presence or absence of office workers, found no significant shift in proportion of uESVs over time irrespective of the presence of office workers. While it was possible to detect the correct burglars’ microbiota as having contributed to the invaded space, the predictions were very weak in comparison to accepted forensic standards. This suggests that at this time 16S rRNA amplicon sequencing of the built environment microbiota cannot be used as a reliable trace evidence standard for criminal investigations.
Keywords: Forensic microbiology | Built-environment | Host-microbe | Trace evidence | Human microbiome
مقاله انگلیسی
89 Delamination analysis using cohesive zone model: A discussion on traction-separation law and mixed-mode criteria
تجزیه و تحلیل لایه لایه شدن با استفاده از مدل منطقه منسجم: بحث در مورد قانون جداسازی کشش و معیارهای حالت مختلط-2020
A discussion on cohesive zone model formulation for prediction of interlaminar damage in composite laminates is presented in this paper. The degradation of interlaminar mechanical properties is analysed from a physical point of view. Firstly, the damage evolution is evaluated according to the traction-separation law and it is demonstrated that if a linear elastic unloading/ reloading curve is assumed, the softening function must also be linear. Secondly, issues regarding damage onset and fracture criteria in mixed-mode loading are critically addressed and commented. A new set of criteria is proposed, and the limitations of existing criteria are discussed.
Keywords: Cohesive zone modelling | Fracture mechanics | Finite element analysis (FEA) | Delamination | Interface fracture
مقاله انگلیسی
90 Dynamic resource allocation during reinforcement learning accounts for ramping and phasic dopamine activity
تخصیص منابع پویا در طول حساب های یادگیری تقویتی برای فعالیت دوپامین ramping و مرحله ای-2020
For an animal to learn about its environment with limited motor and cognitive resources, it should focus its resources on potentially important stimuli. However, too narrow focus is disadvantageous for adaptation to environmental changes. Midbrain dopamine neurons are excited by potentially important stimuli, such as reward-predicting or novel stimuli, and allocate resources to these stimuli by modulating how an animal approaches, exploits, explores, and attends. The current study examined the theoretical possibility that dopamine activity reflects the dynamic allocation of resources for learning. Dopamine activity may transition between two patterns: (1) phasic responses to cues and rewards, and (2) ramping activity arising as the agent approaches the reward. Phasic excitation has been explained by prediction errors generated by experimentally inserted cues. However, when and why dopamine activity transitions between the two patterns remain unknown. By parsimoniously modifying a standard temporal difference (TD) learning model to accommodate a mixed presentation of both experimental and environmental stimuli, we simulated dopamine transitions and compared them with experimental data from four different studies. The results suggested that dopamine transitions from ramping to phasic patterns as the agent focuses its resources on a small number of rewardpredicting stimuli, thus leading to task dimensionality reduction. The opposite occurs when the agent re-distributes its resources to adapt to environmental changes, resulting in task dimensionality expansion. This research elucidates the role of dopamine in a broader context, providing a potential explanation for the diverse repertoire of dopamine activity that cannot be explained solely by prediction error.
Keywords: Prediction error | Salience | Temporal-difference learning model | Pearce-Hall model | Habit | Striatum
مقاله انگلیسی
rss مقالات ترجمه شده rss مقالات انگلیسی rss کتاب های انگلیسی rss مقالات آموزشی
logo-samandehi
بازدید امروز: 8111 :::::::: بازدید دیروز: 0 :::::::: بازدید کل: 8111 :::::::: افراد آنلاین: 64