Industrial forecasting is a top-echelon researchuddomain, which has over the past several years experiencedudhighly provocative research discussions. The scope of thisudresearch domain continues to expand due to the continuousudknowledge ignition motivated by scholars in the area. So,udmore intelligent and intellectual contributions on currentudresearch issues in the accident domain will potentially sparkudmore lively academic, value-added discussions that will be ofudpractical significance to members of the safety community. Inudthis communication, a new grey–fuzzy–Markov time seriesudmodel, developed from nondifferential grey interval analyticaludframework has been presented for the first time. Thisudinstrument forecasts future accident occurrences under timeinvarianceudassumption. The actual contribution made in theudarticle is to recognise accident occurrence patterns anduddecompose theminto grey state principal pattern components.udThe architectural framework of the developed grey–fuzzy–udMarkov pattern recognition (GFMAPR) model has fourudstages: fuzzification, smoothening, defuzzification andudwhitenisation. The results of application of the developedudnovel model signify that forecasting could be effectivelyudcarried out under uncertain conditions and hence, positions the model as a distinctly superior tool for accident forecastingudinvestigations. The novelty of thework lies in the capability ofudthe model inmaking highly accurate predictions and forecastsudbased on the availability of small or incomplete accident data.
展开▼
