Cloud Transformation Algorithm Research Articles

Cloud-VAE: Variational autoencoder with concepts embedded

Variational Autoencoder (VAE) has been widely and successfully used in learning coherent latent representation of data. However, the lack of interpretability in the latent space constructed by the VAE under the prior distribution is still an urgent problem. This paper proposes a VAE with understandable concept embedding named Cloud-VAE, which constructs interpretable latent space by disentangling the latent variables and considering their uncertainty based on cloud model. Firstly, cloud model-based clustering algorithm cast initial constraint of latent space into a prior distribution of concept which can be embedded into the latent space of the VAE to disentangle the latent variables. Secondly, reparameterization trick based on forward cloud transformation algorithm is designed to estimate the latent space concept by increasing the randomness of latent variables. Furthermore, variational lower bound of Cloud-VAE is derived to guide the training process to construct concepts of latent space, realizing the mutual mapping between latent space and concept space. Finally, experimental results on 6 benchmark datasets show that Cloud-VAE has good clustering and reconstruction performance, which can explicitly explain the aggregation process of the model and discover more interpretable disentangled representations.

Cloud-Cluster: An uncertainty clustering algorithm based on cloud model

As a cornerstone of the world, uncertainty embodies the nature of data and knowledge. Existing uncertainty theory-based clustering algorithms learn fuzziness, i.e., the uncertainty of clustering objects belonging to different clusters. However, these algorithms do not refer to the fuzziness of objects themselves, i.e., the randomness of data. Here, we propose a clustering algorithm named Cloud-Cluster, which simultaneously characterizes the fuzziness and randomness of objects to reserve uncertain information, and to describe clusters into concepts. It embeds random uncertainty of concepts to extend the data distribution range for better data partitions and gradually constructs accurate concepts by an improved backward cloud transformation algorithm (MBCT-SR-Ex). Moreover, to ensure that the concept clustering process gradually converges, Cloud-Cluster introduces the Cluster Concept Drift Degree to evaluate the uncertainty of concepts during the clustering process. Experiments on UCI and OpenML clustering datasets show that Cloud-Cluster improves the average clustering accuracy by over 14% compared to K-Means and uncertainty theory-based clustering algorithms. Extensive experimental results on the evaluation of uncertainty show that Cloud-Cluster can handle the uncertainty of datasets in the clustering process well, in addition to exhibiting robustness with unclear clusters.

Cloud transformation algorithm and Copulas function coupling model for drought hazard comprehensive evaluation

Drought hazard is the initial inputting element of regional drought disaster risk management system, which will result in serious drought disaster loss and ecological degradation in combination with the interaction of drought disaster vulnerability factors including population, social economy and ecological environment. Additionally, drought hazard comprehensive evaluation and its variation characteristic analysis are of significant importance for implementing drought disaster resistance schemes. Therefore, given the advantages of Multi-dimensional Precondition Cloud Algorithm (MPCA) in quantifying the uncertainties of drought hazard system and Cloud Transformation Algorithm (CTA) in multivariate frequency distribution derivation, the MPCA, CTA and Copulas function were innovatively introduced to construct drought hazard comprehensive evaluation model (CTCF) aiming to conduct historical variation characteristic exploration and future drought hazard evolution trend prediction analysis. And eventually, the proposed CTCF model was further verified through its application in Anhui Province, China, and the results revealed that, (1) the frequency curve fitting result of drought hazard indicators basing on CTCF approach was effective, which had lower calculation error of RMSE as compared to the corresponding calculation results utilizing traditional normal distribution hypothesis. (2) light-type drought hazard event (Grade II) is the dominant type for the entire provincial area in the future, and the overall occurring probability of future drought hazard events will present gradually declining trend from north to south. The above results had better consistent variation properties with regional historical observed statistical data series of drought disaster system, which indicates that the proposed CTCF model is reasonable in the application of drought hazard comprehensive evaluation field, and will also be beneficial for the implement of drought disaster precaution strategies and construction of ecological environment.

An integrated risk assessment methodology based on fuzzy TOPSIS and cloud inference for urban polyethylene gas pipelines

Due to the superior performance of polyethylene (PE) gas pipes, an increasing number of PE pipelines are being used in urban areas. However, PE gas pipelines are basically located in densely populated areas with high consequences. Once a leak occurs, combustion and explosion accidents can easily happen causing a major threat to human life and property. Furthermore, the assessment results of the common risk assessment methods often have great uncertainty and deviation. It is difficult to accurately identify the main indicators that need priority management. To address these problems, this paper proposes a novel methodology by integrating a fuzzy TOPSIS model and cloud inference for conducting risk assessments of urban PE gas pipelines. The developed methodology is composed of two branches. In the first branch, a rapid risk assessment model for PE pipelines is established to determine whether there are some special emergency situations, which can be directly identified as high risk. In the second branch, a detailed risk assessment process is adopted to obtain the final result, which mainly includes risk factor analysis, risk status assessment and risk classification. Among them, risk factor analysis is used to establish a risk index system and develop a risk index management strategy by means of mathematical statistics and the fuzzy TOPSIS analysis method. The risk status assessment is used to determine the final risk value based on the virtual cloud and fault tree analysis, which is mainly realized by the triangular fuzzy number, structural entropy weight method and backward cloud transformation algorithm. Finally, the risk assessment results of PE gas pipelines are clearly presented in the form of cloud inference, and the final risk classification is determined. The case study shows that the pipeline in this section is at the medium risk level, and the management of time-related indicators and third-party damage is given priority. The result is consistent with the actual situation, which verifies the effectiveness and practicability of the methodology.

Spatio-Temporal Evolution Analysis of Drought Based on Cloud Transformation Algorithm over Northern Anhui Province.

Drought is one of the most typical and serious natural disasters, which occurs frequently in most of mainland China, and it is crucial to explore the evolution characteristics of drought for developing effective schemes and strategies of drought disaster risk management. Based on the application of Cloud theory in the drought evolution research field, the cloud transformation algorithm, and the conception zooming coupling model was proposed to re-fit the distribution pattern of SPI instead of the Pearson-III distribution. Then the spatio-temporal evolution features of drought were further summarized utilizing the cloud characteristics, average, entropy, and hyper-entropy. Lastly, the application results in Northern Anhui province revealed that the drought condition was the most serious during the period from 1957 to 1970 with the SPI12 index in 49 months being less than −0.5 and 12 months with an extreme drought level. The overall drought intensity varied with the highest certainty level but lowest stability level in winter, but this was opposite in the summer. Moreover, drought hazard would be more significantly intensified along the elevation of latitude in Northern Anhui province. The overall drought hazard in Suzhou and Huaibei were the most serious, which is followed by Bozhou, Bengbu, and Fuyang. Drought intensity in Huainan was the lightest. The exploration results of drought evolution analysis were reasonable and reliable, which would supply an effective decision-making basis for establishing drought risk management strategies.

Using the Cloud-Bayesian Network in Environmental Assessment of Offshore Wind-Farm Siting

Offshore wind energy has become the fastest growing form of renewable energy for the last few years. And the development of offshore wind farms (OWFs) is now characterized by a boom. OWF siting is crucial in the success of wind energy projects. Therefore, this paper aims to introduce intelligent algorithms to improve the siting assessment under conditions of multisource and uncertain information. An optimization macrositing model based on Cloud-Bayesian Network (Cloud-BN) is put forward. We introduce the cloud model and adaptive Gaussian cloud transformation (A-GCT) algorithm to grade indicators and apply BN to achieve nonlinear integration and inference of multi-indicators. Combined with the fuzzy representation of the cloud model and probabilistic reasoning of BN, the proposed model can investigate the most efficient siting areas of OWFs in the North Sea of Europe. The experimental results indicate that the siting accuracy is up to 86.67% with reference to the actual OWF location.

Bidirectional Cognitive Computing Model for Uncertain Concepts

Most intelligent computing models are inspired by various human/natural/social intelligence mechanisms during the past 60 years. Achievements of cognitive science could give much inspiration to artificial intelligence. Cognitive computing is one of the core fields of artificial intelligence. It aims to develop a coherent, unified, universal mechanism inspired by human mind’s capabilities. It is one of the most critical tasks for artificial intelligence researchers to develop advanced cognitive computing models. The human cognition has been researched in many fields. Some uncertain theories are briefly analyzed from the perspective of cognition based on concepts. In classical intelligent information systems, original data are collected from environment at first; usually, useful information is extracted through analyzing the input data then, it is used to solve some problem at last. There is a common characteristic between traditional machine learning, data mining, and knowledge discovery models. That is, knowledge is always transformation from data. From the point of view of granular computing, it is a unidirectional transformation from finer granularity to coarser granularity. Inspired by human’s granular thinking and the cognition law of “global precedence”, the human cognition process is from coarser granularity to finer granularity. Generally speaking, concepts (information and knowledge) in a higher granularity layer would be more uncertain than the ones in a lower granularity layer. A concept in a higher granularity layer would be the abstraction of some objects (data or concepts in a lower granularity layer). Obviously, there is a contradiction between the unidirectional transformation mechanism “from finer granularity to coarser granularity” of traditional intelligent information systems with the global precedence law of human cognition. That is, the human cognition are different the computer cognition for uncertain concept. The human cognition for knowledge (or concept) is based on the intension of concept, while the computing of computer (or machine) is based on the extension. In order to integrate the human cognition of “from coarser to finer” and the computer’s information processing of “from finer to coarser”, a new cognitive computing model, bidirectional cognitive computing model between the intension and extension of uncertain concepts, is proposed. The purpose of the paper is to establish the relationship between the human brain computing mode (computing based on intension of concept) and the machine computing mode (computing based on extension of concept) through the way of computation. The cloud model theory as a new cognition model for uncertainty proposed by Li in 1995 based on probability theory and fuzzy set theory, which provides a way to realize the bidirectional cognitive transformation between qualitative concept and quantitative data—forward cloud transformation and backward cloud transformation. Inspired by the cloud model theory, the realization of the bidirectional cognitive computing process in the proposed method is that the forward cloud transformation algorithm can be used to realize the cognitive transformation from intension to extension of concept, while the backward cloud transformation algorithm is to realize the cognitive transformation from extension to intension. In other words, the forward cloud transformation is a converter “from coarser to finer”, and the backward cloud transformation is a converter “from finer to coarser”. Taking some uncertain concepts as cognitive unit of simulation, several simulation experiments of the bidirectional cognition computing process are implemented in order to simulate the human cognitive process, such as cognition computing process for an uncertain concept with fixed samples, cognition computing process of dynamically giving examples, and cognition computing process of passing a concept among people. These experiment results show the validity and efficiency of the bidirectional cognitive computing model for cognition study.

Risk assessment for long-distance gas pipelines in coal mine gobs based on structure entropy weight method and multi-step backward cloud transformation algorithm based on sampling with replacement

Destruction of pipelines because of coal mine gob disasters may result in enormous financial loss and significantly affect public safety. Hence, risk assessment of gob pipelines is of immense significance for field personnel. Given the lack of statistical data and the limitations of expert experience, a prior single risk value is often insufficient to reflect the actual situation and cannot meet the needs of the site. The backward cloud transformation (BCT) algorithm is a method that can fully mine the local information contained in expert experience to restore the overall information. However, the existing BCT algorithm has no solution under certain conditions, which considerably limits its application. This study proposes a comprehensive risk assessment method by combining the structural entropy method with a multi-step backward cloud transformation algorithm based on sampling with replacement (MBCT-SR). First, a simplified model for rapid identification is used to determine whether it is worth calculating risk values. Second, a fault tree that fits the actual situation is established, and the weights of the indexes are determined by the structural entropy weight method. Third, the interval scores of the indexes are transformed into numerical features of the cloud model, which are then logically operated using virtual cloud algorithms. Finally, the risk values of the pipeline can be obtained, the cloud droplet diagram of the risk values is clearly shown by the forward cloud transformation (FCT) algorithm, and the risk level can be obtained by the probability that the cloud droplet falls into each risk level interval. To validate the utility of the proposed method, a case study of a coal mine gob around a long-distance gas pipeline was investigated.

A novel link quality prediction algorithm for wireless sensor networks

Ahead knowledge of link quality can reduce the energy consumption of wireless sensor networks. In this paper, we propose a cloud reasoning-based link quality prediction algorithm for wireless sensor networks. A large number of link quality samples are collected from different scenarios, and their RSSI, LQI, SNR and PRR parameters are classified by a self-adaptive Gaussian cloud transformation algorithm. Taking the limitation of nodes? resources into consideration, the Apriori algorithm is applied to determine association rules between physical layer and link layer parameters. A cloud reasoning algorithm that considers both short- and long-term time dimensions and current and historical cloud models is then proposed to predict link quality. Compared with the existing window mean exponentially weighted method, the proposed algorithm captures link changes more accurately, facilitating more stable prediction of link quality

A Triangular Cloud Model and Cloud Generator Algorithm

Cloud model is a cognitive model which represents the fuzziness and randomness with three digital characteristics, i.e. expectation, entropy and hyper-entropy. The normal cloud model is one of the most important cloud models, which is composed of 2nd-order normal distribution. The normal cloud model is usually expressed in qualitative concept with symmetrical features, and for the concept of fasymmetric only uses the complex combination of cloud model. Therefore, it is more complex in computing and processing in practical application. A new simple cloud model, i.e. triangular cloud model is proposed, and the forward triangular cloud transformation algorithm is presented. Then, a backward triangular cloud transformation algorithm is designed based on the fuzzy discrete expectation. Finally, two simulation examples are given to state the feasibility of the triangular cloud, and the result shows that the new generator algorithm is simple and effective.

A novel hybrid water quality time series prediction method based on cloud model and fuzzy forecasting

Accurate water quality time series prediction can provide support to early warning of water pollution as well as decision-making for water resource management. Due to the uncertainty of the water quality data including randomness, fuzziness, imprecision, and nonstationary, the prediction accuracy of the traditional models has been limited. In this paper, a multi-factor water quality time series prediction model is proposed, based on Heuristic Gaussian cloud transformation, the approximate periodicity of water quality parameter and fuzzy time series model. The proposed model uses the Heuristic Gaussian cloud transformation algorithm to extract the uncertain numerical time series into Gaussian clouds, and constructs the training dataset by calculating the length of the approximate periodicity, which can greatly reduce the noise data. Then, it applies the fuzzy time series model to do the prediction. The proposed model is tested for DO, CODMn, water temperature and EC prediction. The experimental results show that the proposed method significantly improved the prediction accuracy compared with the existing time series prediction models for water quality prediction.

Generic normal cloud model

Cloud model is a cognitive model which can realize the bidirectional cognitive transformation between qualitative concept and quantitative data based on probability statistics and fuzzy set theory. It uses the forward cloud transformation (FCT) and the backward cloud transformation (BCT) to implement the cognitive transformations between the intension and extension of a concept. As one of the most important cloud models, the normal cloud models, especially the 2nd-order normal cloud model based on normal distribution and Gaussian membership function has been extensively researched and successfully applied to many fields. In this paper, a 2nd-order generic normal cloud model, which establishes a relationship between normal cloud and normal distribution, is proposed, and the 2nd-order generic forward normal cloud transformation algorithm (2nd-GFCT) is presented. Whereafter, an ideal backward cloud transformation algorithm of the 2nd-order generic normal cloud model (2nd-GIBCT) is designed based on the mutually inverse features of FCT and BCT, in which the distribution of all the cloud drops generated in 2nd-GFCT is used. Meanwhile, a 2nd-order generic backward cloud transformation algorithm (2nd-GBCT), which does not use the distribution of cloud drops, is also proposed to solve real life problems since it is impossible to know the distribution of all the cloud drops in advance in real life applications. The relationships between the generic backward cloud transformation algorithms are further studied, which help reach the finding that the two backward cloud transformation algorithms presented by Wang and Xu [26,34] are two special cases of the 2nd-GBCT. In addition, the 2nd-order generic normal cloud model is further generalized to pth-order generic normal cloud model, and the pth-order generic forward normal cloud transformation algorithm (pth-GFCT) and the backward cloud transformation algorithm (pth-GBCT) are presented. Finally, the performances of the 2nd-GIBCT and the 2nd-GBCT are illustrated by simulation experiment. The effectiveness of the 2nd-GBCT is shown by the results of image segmentation.

A New Multi-Step Backward Cloud Transformation Algorithm Based on Normal Cloud Model

The representation and processing of uncertainty information is one of the key basic issues of the intelligent information processing in the face of growing vast information, especially in the era of network. There have been many theories, such as probability statistics, evidence theory, fuzzy set, rough set, cloud model, etc., to deal with uncertainty information from different perspectives, and they have been applied into obtaining the rules and knowledge from amount of data, for example, data mining, knowledge discovery, machine learning, expert system, etc. Simply, This is a cognitive transformation process from data to knowledge (FDtoK). However, the cognitive transformation process from knowledge to data (FKtoD) is what often happens in human brain, but it is lack of research. As an effective cognition model, cloud model provides a cognitive transformation way to realize both processes of FDtoK and FKtoD via forward cloud transformation (FCT) and backward cloud transformation (BCT). In this paper, the authors introduce the FCT and BCT firstly, and make a depth analysis for the two existing single-step BCT algorithms. We find that these two BCT algorithms lack stability and sometimes are invalid. For this reason we propose a new multi-step backward cloud transformation algorithm based on sampling with replacement (MBCT-SR) which is more precise than the existing methods. Furthermore, the effectiveness and convergence of new method is analyzed in detail, and how to set the parameters m, r appeared in MBCT-SR is also analyzed. Finally, we have error analysis and comparison to demonstrate the efficiency of the proposed backward cloud transformation algorithm for some simulation experiments.