Relevant Input Factors Research Articles

Partially latent factors based multi‐view subspace learning

AbstractMulti‐view subspace clustering always performs well in high‐dimensional data analysis, but is sensitive to the quality of data representation. To this end, a two‐stage fusion strategy is proposed to embed representation learning into the process of multi‐view subspace clustering. This article first proposes a novel matrix factorization method that can separate the coupling consistent and complementary information from observations of multiple views. Based on the obtained latent representations, we further propose two subspace clustering strategies: feature‐level fusion and subspace‐level hierarchical strategy. The feature‐level method concatenates all kinds of latent representations from multiple views, and the original problem therefore degenerates to a single‐view subspace clustering process. The subspace‐level hierarchical method performs different self‐expressive reconstruction processes on the corresponding complementary and consistent latent representations coming from each view, that is, the prior constraints imposed on different types of subspace representations are related to the relevant input factors. Finally, extensive experimental results on real‐world datasets demonstrate the superiority of our proposed methods by comparing them against some state‐of‐the‐art subspace clustering algorithms.

Probabilistic Residential Load Forecasting Based on Micrometeorological Data and Customer Consumption Pattern

A prior knowledge of residential load demand is critical for power system operations at the distribution level, such as economic dispatch, demand response and energy storage schedule. However, as residential customers perform more casual and active consumption behaviors, prediction of such highly volatile loads can be much harder. Owing to the development of sensor technology, micrometeorological data can be sampled with a high geographic resolution. Those data that represent the weather condition on the land surface show a strong relationship to the residential load evidently, whereas it remains unsolved on how to fully utilize those great number of datasets. This paper proposes a day-ahead probabilistic residential load forecasting method based on a novel deep learning model, named convolutional neural network with squeeze-and-excitation modules (CNN-SE), and micrometeorological data. The model can employ multi-channel input data with dissimilar weights, suitable for analyzing massive relevant input factors. A feature extraction method is adopted for customer consumption pattern based on sparse auto-encoder (SAE), which can help correct probabilistic forecasting results. A case study that covers 8 residential communities and 18 micrometeorological sites is conducted to validate the feasibility and accuracy of the proposed hybrid method.

Challenges for the European steel industry: Analysis, possible consequences and impacts on sustainable development

The steel industry in the European Union (EU), important for the economy as a whole, faces various challenges. These are inter alia volatile prices for relevant input factors, uncertainties concerning the regulation of CO2-emissions and market shocks caused by the recently introduced additional import duties in the US, which is an important sales market. We examine primary and secondary effects of these challenges on the steel industry in the EU and their impacts on European and global level. Developing and using a suitable meta-model, we analyze the competitiveness of key steel producing countries with respect to floor prices depending on selected cost factors and draw conclusions on the impacts in the trade of steel on emissions, energy demand, on the involvement of developing countries in the value chain as well on the need for innovations to avoid relocations of production. Hence, our study contributes to the assessment of sustainable industrial development, which is aimed by the Sustainability Development Goal “Build resilient infrastructure, promote inclusive and sustainable industrialization and foster innovation countries”. By applying information on country-specific Human Development Indexes (reflecting aspects of life expectancy, education, and per capita income), we show that relocating energy-intensive industries from the EU may not only increase global energy demand and CO2-emissions, but may also be to the disadvantage of developing countries.

Learning Dynamic Factors to Improve the Accuracy of Bus Arrival Time Prediction via a Recurrent Neural Network

Accurate prediction of bus arrival times is a challenging problem in the public transportation field. Previous studies have shown that to improve prediction accuracy, more heterogeneous measurements provide better results. So what other factors should be added into the prediction model? Traditional prediction methods mainly use the arrival time and the distance between stations, but do not make full use of dynamic factors such as passenger number, dwell time, bus driving efficiency, etc. We propose a novel approach that takes full advantage of dynamic factors. Our approach is based on a Recurrent Neural Network (RNN). The experimental results indicate that a variety of prediction algorithms (such as Support Vector Machine, Kalman filter, Multilayer Perceptron, and RNN) have significantly improved performance after using dynamic factors. Further, we introduce RNN with an attention mechanism to adaptively select the most relevant input factors. Experiments demonstrate that the prediction accuracy of RNN with an attention mechanism is better than RNN with no attention mechanism when there are heterogeneous input factors. The experimental results show the superior performances of our approach on the data set provided by Jinan Public Transportation Corporation.

Application of Welding Process Parameters Using AI Algorithm

Welding is a complex process that plays a critical role in various industries, including manufacturing, construction, and automotive. Achieving optimal welding process parameters is essential to ensure high-quality welds with minimal defects. However, determining the optimal parameters requires extensive experimentation and expertise, which can be time-consuming and costly. In artificial intelligence (AI) algorithms have shown great potential in optimizing welding process parameters. This study focuses on the application of AI algorithms for optimizing welding process parameters. The objective is to develop a machine learning model that can accurately predict the optimal welding parameters based on various input factors such as material type, joint configuration, and desired weld quality. The model utilizes a dataset of past welding experiments, where the parameters and corresponding weld quality were recorded. The significance of this research lies in its potential to revolutionize the welding industry by improving the efficiency and effectiveness of the welding process. By leveraging AI algorithms, the model can quickly analyze and identify the relationships between input factors and welding parameters, leading to optimal parameter recommendations. This can result in reduced welding defects, improved weld quality, and increased productivity.
 To achieve the research objectives, a comprehensive literature review is conducted to understand the current state of AI algorithms applied to welding process optimization. Various AI techniques such as neural networks, genetic algorithms, and support vector machines are explored for their suitability in this context. The literature review also covers previous studies and their findings on the application of AI algorithms in welding parameter optimization. The methodology of this research involves the collection of welding process data from different welding experiments. This dataset is then pre- processed to handle missing values, outliers, and normalize the data. Feature engineering techniques are applied to select the most relevant input factors. Several AI algorithms are implemented and trained using the dataset, and their performance is evaluated using appropriate metrics such as accuracy and mean squared error. The accuracy of parameter predictions and the effectiveness of the optimized welding process parameters are evaluated. Furthermore, the implications of using AI algorithms for welding parameter optimization are discussed, highlighting the potential benefits and limitations of this approach. This research demonstrates the applicability and effectiveness of AI algorithms in optimizing welding process parameters. The developed machine learning model offers a promising solution to expedite the parameter optimization process, leading to improved weld quality and reduced costs. The practical implications of this study include increased productivity, enhanced weld quality control, and improved overall efficiency in the welding industry. The findings of this research contribute to the growing field of AI-assisted welding process optimization and pave the way for further advancements in this area.

Static total cost comparison model to identify economic fields of application of Cellular Manufacturing for milling and drilling processes versus done-in-one-concepts

The predominant approach on machining is to manufacture a workpiece on a complex done-in-one machine tool. In contrast, the Cellular Manufacturing concept splits up the machining steps and distributes them sequentially to several simpler and cost-efficient machine tools with a higher manual workload. Both concepts are technically feasible, but the question which approach is favorable from an economic perspective is still open.This article first identifies relevant input factors like machining tasks, operator work content, labor and capital costs, etc. Available machine tools are grouped by a k-means cluster analysis, human tasks of operation are identified by an MTM-UAS analysis. Together with key performance indicators like ‘cell balancing efficiency’, all data are transferred into monetary values and combined to a static total cost comparison model. The paper closes with a validation on industrial parts and a sensitivity analysis of the results achieved.

Modeling airborne benzene in space and time with self-organizing maps and Bayesian techniques

An assessment of personal exposure to airborne chemical contaminants demands for individual-specific registration of their concentrations, a procedure which is expensive and difficult to implement. An alternative approach is the calculation of a spatial concentration field in high resolution where exposure can be assigned to individuals according to their dwelling locations. Self-organizing maps (SOM) and Bayesian Hierarchical Models (BHM) were applied to model the spatial concentrations of benzene, an airborne volatile organic compound (VOC), in the urban area of Leipzig, Germany. Different performance measures (mean absolute error, coefficient of determination, etc.) were adopted to evaluate and compare the performance of SOM and BHM. Relevant input factors related to VOC dispersion were stepwise selected with the BHM. Both modeling techniques identified seasonal as well as spatial variations of benzene, with the highest concentrations occurring in winter and the lowest in summer. SOM and BHM showed that high concentrations of benzene are correlated with low distances to the city center and with the major traffic routes. Both SOM and BHM were suitable to model the spatial distribution of benzene concentrations, with the latter yielding a better overall performance using input factors selected by BHM. Beyond this specific application the suggested approaches have potential for statistical spatiotemporal modeling of other environmental parameters, an issue that is currently under rapid development.

Global uncertainty and sensitivity analysis of a food‐web bioaccumulation model

A global uncertainty and sensitivity analysis (UA/SA) of a state-of-the-art, food-web bioaccumulation model was carried out. We used an efficient screening analysis technique to identify the subset of the most relevant input factors among the whole set of 227 model parameters. A quantitative UA/SA was then applied to this subset to rank the relevance of the parameters and to partition the variance of the model output among them by means of a nonlinear regression of the outcomes of 1,000 Monte Carlo simulations. The concentrations of four representative persistent organic pollutants (POPs) in two representative species of the coastal marine food web of the Lagoon of Venice (Italy) were taken as model outputs. The screening analysis showed that the ranking was remarkably different in relation to the species and chemical being considered. The subsequent Monte Carlo-based quantitative analysis pointed out that the relationships among some of the parameters and the model outputs were nonlinear. The nonlinear regression showed that the fraction of output variance accounted for by each parameter was strongly dependent on the range of the octanol-water partition coefficient (K(OW)) values being considered. For the less hydrophobic chemicals, the main sources of model uncertainty were the parameters related to the respiratory bioaccumulation, whereas for the more hydrophobic ones, K(OW) and the other parameters related to the dietary uptake explained the largest fractions of the variance of the chemical concentrations in the organisms. The analysis highlighted that efforts are still needed for reducing uncertainty of model parameters to get reliable results from the application of food web bioaccumulation models.

Inequity in Hindu–Muslim Riots: A Test of Two Biases1

During the past 50 years, substantially more Muslims than Hindus have died in riots in India. This study examined and found support for 2 possible individual‐level reasons that are consistent with this disproportionate number of deaths. First, Hindus, compared to Muslims, considered historical incidents to be more important for contemporary Hindu–Muslim relations. Second, Hindus showed an ultimate group bias in that they generally valued Hindu lives more than Muslim lives, whereas Muslims showed no difference in the valuation of Hindu and Muslim lives. Together, these results are consistent with an equity notion that Hindus consider as relevant input factors both historical incidents and perceptions that in‐group members’ lives are more valuable than are out‐group members’ lives.

Time series forecasting using backpropagation neural networks

This paper describes a neural network approach for time series forecasting. This approach has several significant advantages over other conventional forecasting methods such as regression and Box-Jenkins; besides simplicity, another major advantage is that it does not require any assumption to be made about the underlying function or model to be used. All it needs are the historical data of the target and those relevant input factors for training the network. In some cases, even the historical targets alone are sufficient to train the network for forecasting. Once the network is well trained and the error between the target and the network forecasts has converged to an acceptable level, it is ready for use. The proposed network has a three-dimensional structure which is proposed for capturing the temporal information contained in the input time series. Several real applications, including forecasting of electricity load, stock market and interbank interest rate forecastings were tested with the proposed network and the findings were very encouraging.

Sampling Procedures for Risk Simulation

Risk simulation has become a standard alternative to the utilization of single estimates for input data to investment appraisal models. Under the risk simulation approach managers define probability distributions for relevant input factors. During this process management may consciously or subconsciously assume certain relationships between the input variables. These interdependencies are often omitted from the models, however, under the justification that they are accounted for implicitly in the values attributed to the variables by management. It is demonstrated in this paper that such omissions may lead to significant errors in the observed distributions of the selected appraisal criteria. Various forms of discriminate sampling are introduced and an example is cited for which certain discriminate schemes are shown to be more accurate than traditional independent sampling.