Multivariate data analysis

Multivariate data analysis

Multivariate data analysis gaining its momentum as Mahalanobis Taguchi system is helpful in solving many issues over the period. MTS has developed from both Mahalanobis distance (MD) and Taguchi’s Robust Engineering [8]. It has used for decision making in recent days under multi-attribute criteria. This system is a multivariate statistical method used for feature selection as well [9]. Survey of MTS was presented, which illustrates how it could produce better results for feature selection as well as for classification problems [10]. Better Arrhythmia classification system was developed using MTS, which has a better classification result compared to the other well-known approaches [11].

Imbalanced data classification was successfully handled with the help of a modified MTS approach. The improved version’s results bettered not only the MTS approach but also SVM and Naïve Bayes approach and many other algorithms [12]. Smartphone sensed data was used to develop a model to assess the quality of the road was proposed with the help of MTS approach. Road condition classification was handled in this work using MTS [13]. In another work, human gaits are classified by using the MTS algorithm. It was done to assess the ACLR patients and the proposed approach has produced better results [14]. Screening and ranking of health hazard were done with the help of MTS algorithm in another work in which correlation between the variables was considered for this ranking framework [15]. Better results were achieved in this model compared to the other familiar ranking frameworks.

Bearing fault diagnosis was made in this work under variable conditions. This system was executed with the help of MTS approach. In industry, this could be a typical application as the monitoring of all these manually could be a difficult task and this work has automated that. In this work, the MTS approach was used for the feature selection purpose [16]. A modified MTS called MTGS was proposed to evaluate the energy security performance. It was tested in china, where the collected data from 2013 to 2017 were considered. Here the improved MTS was used for weights calculation [17].

On the other side, the grey correlation method is a powerful approach which could be applied to different types of problem-solving scenario. Recently the multi-objective optimization was done with the help of this approach in duplex stainless steel cutting force analysis [18]. Assessment of Dam disaster status was analyzed using grey relation analysis which has produced better outcomes compared to other approaches. In this hybrid approach, this one is used for the weight calculation process [19]. Design methodology for photolithography process was proposed using GRA. A large area was considered for this proposed work and optimal design parameters were obtained in this work using GRA method [20].

In one another work GRA approach was used to calculate the friction signals’ coefficient. In this work, characteristic parameter dG was obtained with the help of this approach [21]. Better results were produced with this approach. Taguchi based GRA was proposed for multi-factor optimization. Here in this work, optimal parametric combination to find the appropriate tensile strength. Improvement of about 10% is achieved with the help of this approach [22]. The interval-valued triangular fuzzy approach has used along with GRA was used for enhancing green supply management process. Critical aspects and criteria are analyzed in this work using the proposed approach to achieve this [23].

A hybrid approach was proposed with the help of fuzzy and GRA for car selection. In the modern era, with the considerable availability of manufacturers and models, this could help in choosing the right car based on one’s need. Both qualitative and quantitative parameters were considered for evaluation and among that GRA was used for identifying quantitative criteria [24]. Ensemble model for clustering was proposed with the help of was proposed. This work was aimed to improve the accuracy of the results of clustering substantially and GRA method was considered for this ensemble model. Also, this proposed approach has overcome the category label matching problem, which generally presents in the traditional approaches. Grey linked matrix are produced here from the base clusters to overcome this issue [25].