摘要 :
We elucidate the practical implementation of Spiking Neural Network (SNN) as local ensembles of classifiers. Synaptic time constant rs is used as learning parameter in representing the variations learned from a set of training dat...
展开
We elucidate the practical implementation of Spiking Neural Network (SNN) as local ensembles of classifiers. Synaptic time constant rs is used as learning parameter in representing the variations learned from a set of training data at classifier level. This classifier uses coincidence detection (CD) strategy trained in supervised manner using a novel supervised learning method called rs Prediction which adjusts the precise timing of output spikes towards the desired spike timing through iterative adaptation of τ_s. This paper also discusses the approximation of spike timing in Spike Response Model (SRM) for the purpose of coincidence detection. This process significantly speeds up the whole process of learning and classification. Performance evaluations with face datasets such as AR, FERET, JAFFE, and CK+ datasets show that the proposed method delivers better face classification performance than the network trained with Supervised Synaptic-Time Dependent Plasticity (STDP). We also found that the proposed method delivers better classification accuracy than k nearest neighbor, ensembles of kNN, and Support Vector Machines. Evaluation on several types of spike codings also reveals that latency coding delivers the best result for face classification as well as for classification of other multivariate datasets.
收起
摘要 :
This article describes a systematic approach to subgroup classification based on a classification framework and sequential steps involved in the subgrouping process. The sequential steps are stating the purpose of the classificati...
展开
This article describes a systematic approach to subgroup classification based on a classification framework and sequential steps involved in the subgrouping process. The sequential steps are stating the purpose of the classification, identifying the classification elements, using relevant information, and using clearly stated and purposeful subgroup classification terms. This systematic approach reflects current changes in the field of intellectual disability (ID), the modern and social understanding of ID, and the multiple purposes for classification.
收起
摘要 :
Purpose - The purpose of this study is to propose a taxonomy of meeting purpose. Meetings are a workplace activity that deserves increased attention from researchers and practitioners. Previous researchers attempted to develop typ...
展开
Purpose - The purpose of this study is to propose a taxonomy of meeting purpose. Meetings are a workplace activity that deserves increased attention from researchers and practitioners. Previous researchers attempted to develop typologies of meeting purpose with limited success. Through a comparison of classification methodologies, the authors consider a taxonomy as the appropriate classification scheme for meeting purpose. The authors then utilize the developed taxonomy to investigate the frequency with which a representative sample of working adults engaged in meetings of these varying purposes. Their proposed taxonomy provides relevant classifications for future research on meetings as well and serves as a useful tool for managers seeking to use and evaluate the effectiveness of meetings within their organizations. Design/methodology/approach - This study employs an inductive methodology using discourse analysis of qualitative meeting descriptions to develop a taxonomy of meeting purpose. The authors discourse analysis utilizes open-ended survey responses from a sample of working adults (n = 491). Findings - The authors categorical analysis of open-ended questions resulted in a 16-category taxonomy of meeting purpose. The two most prevalent meeting purpose categories in this sample were "to discuss ongoing projects" at 11.6 per cent and "to routinely discuss the state of the business" at 10.8 per cent. The two least common meeting purpose categories in this sample were "to brainstorm for ideas or solutions" at 3.3 per cent and "to discuss productivity and efficiencies" at 3.7 per cent. The taxonomy was analyzed across organizational type and employee job level to identify differences between those important organizational and employee characteristics. Research limitations/implications - The data suggested that meetings were institutionalized in organizations, making them useful at identifying differences between organizations as well as differences in employees in terms of scope of responsibility. Researchers and managers should consider the purposes for which they call meetings and how that manifests their overarching organizational focus, structure and goals. Originality/value - This is the first study to overtly attempt to categorize the various purposes for which meetings are held. Further, this study develops a taxonomy of meeting purposes that will prove useful for investigating the different types of meeting purposes in a broad range of organizational types and structures.
收起
摘要 :
Conventional soil survey information is often unclear except to specialists and experienced soil experts who are in short supply to meet user demands. An approach using a special purpose soil identification key and conceptual topo...
展开
Conventional soil survey information is often unclear except to specialists and experienced soil experts who are in short supply to meet user demands. An approach using a special purpose soil identification key and conceptual toposequence models was developed to assist non-soil experts with identifying soil types in Kuwait. The approach supports the restoration of Kuwait rangelands, where there is a need to assist revegetation success by removing uncertainty regarding soil conditions and targeting the planting of appropriate vegetation communities to the soil type. Legacy data from soil survey reports were available for reinterpretation. The soil identification key developed is in a matrix form and allowed soil types to be determined by the presence or absence of three recognizable soil features, which generally typify arid zone soils worldwide: hardpan, gypsum, and calcium carbonate. The soil type categories are descriptively named for ease of understanding by non-technical users, and were structured to align with the previously identified Soil Taxonomy classes to maintain linkages with the soil survey and other interpreted information. To complement the soil identification key, conceptual soil toposequence models presented the general soil distribution patterns in a visual format to aid understanding of spatial variation and soil type relationships. The approach is flexible and can be scaled with additional criteria as more knowledge is acquired regarding the relationship between soil types and vegetation communities, and while the detail is applicable to Kuwait the approach could be adjusted and applied elsewhere.
收起
摘要 :
With the networking of urban rail transit and the large‐scale development of bikesharing,
metro and bike‐sharing connection has become the preferred way of daily travel
for residents of Xiamen. Current studies mainly identify ...
展开
With the networking of urban rail transit and the large‐scale development of bikesharing,
metro and bike‐sharing connection has become the preferred way of daily travel
for residents of Xiamen. Current studies mainly identify metro station types based on
node and place orientation, lacking behaviour‐based investigation. To fill this gap, this
study aims to explore the classification of metro stations based on transfer purposes by
combining bike‐sharing and point of interest data in Xiamen, using buffer analysis, kernel
density estimation, and DBSCAN clustering algorithm comprehensively. The results
indicate the following. (1) Distinct transfer purposes have significant agglomeration
characteristics and present poly‐centric spatial pattern, an authentic portrayal of
Xiamen's land use function. (2) The heterogeneity of connection flow between different
transfer purposes and metro stations is apparent. The distribution of flow and flow
direction within the same transfer purpose is also in non‐equilibrium. (3) Based on traffic
connection analysis, metro stations are divided into seven types: transportation hub,
employment‐oriented, residence‐oriented, job‐housing balance, school‐oriented, traffictourism
integration, and business connection types. The obtained results assist in
improving the transportation connection environment, perfecting urban land use
planning, and enhancing low‐carbon and green travel.
收起
摘要 :
The generation of substantial amounts of travel- and mobility-related data has spawned the emergence of the era of big data. However, this data generally lacks activity-travel information such as trip purpose. This deficiency led ...
展开
The generation of substantial amounts of travel- and mobility-related data has spawned the emergence of the era of big data. However, this data generally lacks activity-travel information such as trip purpose. This deficiency led to the development of trip purpose inference (activity type imputation/annotation) techniques, of which the performance depends on the available input data and the (number of) activity type classes to infer. Aggregating activity types strongly increases the inference accuracy and is usually left to the discretion of the researcher. As this is open for interpretation, it undermines the reported inference accuracy. This study developed an optimised classification methodology by identifying classes of activity types with an optimal balance between improving model accuracy, and preserving activity information from the original data set. A sensitivity analysis was performed. Additionally, several machine learning algorithms are experimented with. The proposed method may be applied to any study area.
收起
摘要 :
In a general manner, customer relationship management engages in understanding customer needs and meet them. Most of the investments are either far from customer needs or based on a primitive data collection method. However, custo...
展开
In a general manner, customer relationship management engages in understanding customer needs and meet them. Most of the investments are either far from customer needs or based on a primitive data collection method. However, customers mainly do not behave with the same ideas to shop in the retail domain. Several studies aim to understand the visiting purposes of customers using various methods. This study seeks to uncover the visit purposes of customers from their paths. Due to customers' unpredictable moods and plenty of stores in the shopping mall, the discovered paths are usually too complicated to analyze. Process mining that can overcome this obstacle is a method that creates process flows from event logs in the databases. In this study, the visited stores were seen as an activity in a business process. PALIA, a discovery algorithm in process mining, was applied to find and cluster customer paths. This study contributes to the literature by examining customer needs from their indoor paths, which were created by the PALIA algorithm. It facilitates to analyze discrepancies among the visits for the same customer. Moreover, the discovered paths are considered according to the age groups predicted by Levenshtein fuzzy kNN (L-FkNN).
收起
摘要 :
In this study, the drinking groundwater quality of Lenjanat plain, Iran, is classified based on water quality index (WQI), Takagi-Sugeno-Kang fuzzy water quality index (TSKFWQI) and entropy weighted water quality index (EWQI). Gro...
展开
In this study, the drinking groundwater quality of Lenjanat plain, Iran, is classified based on water quality index (WQI), Takagi-Sugeno-Kang fuzzy water quality index (TSKFWQI) and entropy weighted water quality index (EWQI). Groundwater samples from 79 regional monitoring wells and different resources such as agricultural and potable deep wells, rural dug wells, industrial and recreational facilities and drilled wells in the vicinity of pollution sources of urban and rural sewage discharge points were collected and analyzed during 2009-2010. In this research, physicochemical parameters including As, Pb, Cr, Ni, Cu, NO3, Na, K, F, Cl, Ba, Ca, Mg, Fe, SO4 and TDS were used to calculate the drinking quality rank of water samples using WQI, TSKFWQI and EWQI methods. Calculations showed that ranking the groundwater samples using WQI is very similar to ranks determined by entropy-based calculations of water quality index, while the TSKFWQI clearly indicates that this classification method acts stricter than two other methods (WQI and EWQI). In TSKFWQI, the final rank of any sample is very much affected by toxic parameters. It means that a sample with acceptable range of all parameters, except one toxic parameter, falls in the unacceptable rank. As a result, in areas where water chemistry shows the presence of some toxic elements in the groundwater resources, TSKFWQI classification of water with regard to drinking purposes gives more reliable results.
收起
摘要 :
One of the most important agricultural areas in Iran is Golestan Province. In this study, the suitability of groundwater for irrigation purposes was assessed. To determine hydro-geochemical classification of the station and assess...
展开
One of the most important agricultural areas in Iran is Golestan Province. In this study, the suitability of groundwater for irrigation purposes was assessed. To determine hydro-geochemical classification of the station and assessment of spatial modeling of the study area, 12 groundwater stations were selected and an average of the last 5 years of sampling (2011-2015) has been considered. Suitability of groundwater for irrigation was assessed based on Sodium Adsorption Ratio (SAR), Residual Sodium Carbonate (RSC), Na%, Permeability Index (PI), Magnesium Adsorption Ratio (MAR), Kelley's Ratio (KR), Corrosion Ratio (CR), Residual Sodium Carbonate (RSC), Potential Soil Salinity (PSS), Chloro-Alkaline Indices (CAI), Meteoric genesis (Met. gen) and Wilcox classification. The abundance of major ions followed a HCO3 > SO4 > Na > Ca > Mg > Cl > K trend. CAI results indicate reverse cation exchange for this area, and Wilcox classification indicates that only Kalou, Kafshgari, and Kia stations are appropriate for irrigation. The spatial distribution of groundwater quality demonstrates a decreasing trend from SE to NW based on CAI, Cl, CR, EC, TSS, and TDS and shows reverse trends for KR, MAR, Met. gen, Mg/Ca, Na%, pH, PI, and RSC. In addition, the results of hydro-geochemical classification of irrigation parameters have categorized the irrigation water of this area into three different classes. Based on EC, MAR, Mg/Ca, Cl and CR, this area is suitable, but based on PSS, KR, and TH, all stations are unsuitable and according to SAR, TDS, Na% and PI, this area's groundwater is moderate.
收起
摘要 :
This work introduces a methodology to find solutions corresponding to different purposes in a multiple attribute decisionmaking problem under fuzzy environment. The discernment of purpose-based solutions becomes important when the...
展开
This work introduces a methodology to find solutions corresponding to different purposes in a multiple attribute decisionmaking problem under fuzzy environment. The discernment of purpose-based solutions becomes important when the problem is defined vaguely and solution is targeted to heterogeneous population. Depending on the purpose, for which the solution is sought, the attributes are identified and weighted in an appropriate proportion. The level of similarity between a pair of attributes plays an important role to determine the aggregated value of attributes specific to a purpose. Our work determines the similarity levels between a pair of attributes by calculating their maximum attainability in presence of each other. The achievement of an attribute in presence of another is represented as a fuzzy set in the unit interval. The crisp equivalents of the fuzzy sets in the unit interval are used to define their simultaneous satisfaction denoted as 1-step relation. The 1-step relation is extended to (m-1)-step relation to calculate the degree of attainability of the same pair of attributes in the presence of m (all) attributes. The different levels of (m-1)-step relations generate several partitions of the attributes corresponding to multiple purposes in the multiple attribute decision-making problems. The degree of fulfilment of the purposes in the alternatives are numerically derived by first taking weighted average of attributes within the equivalence classes of a partition and then aggregating the values corresponding to equivalence classes through ordered weighted averaging. The methodology is illustrated with a numerical example.
收起