Cloud Computing (CC) is a form of internet-based computing in which shared, configurable resources are made accessible to computers and other devices as a service. With the increasing popularity of CC, the development of high-quality cloud applications has become a crucial area of research. CC is an architectural style that provides cloud users with on-demand or pay-per-use access to a shared pool of services and computational resources [1,2]. Consumers and businesses can save money on capital expenditures and operating expenses owing to CC. It offers users a network-based environment that enables the sharing of computations and services independent of location [3].

With the increasing popularity of service-computing technologies, cloud services have emerged as essential resources that are commonly employed in the context of software systems. Using a pay-as-you-go concept, customers can select services that are stored in cloud data centers [4]. When selecting a service, users usually prioritize their own unique functional needs. With the increasing scope of services offered by cloud data centers, there may be multiple potential services that meet the customer’s requirements. Due to the fact that these candidate services may all match the functional requirements, the user may be required to evaluate them from a non-functional perspective (considering characteristics such as availability, security, reliability, and cost, etc.) before making a choice. Among all the non-functional indicators of cloud services, reliability is arguably the most important and should thus be the primary factor when selecting services [5,6,7].

CC is a business model that enables convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with little management effort or interaction from service providers [8]. Through virtualization, CC can support a huge customer base with varying computation needs on the same physical infrastructure. Contrary to prior paradigms, such as clusters and grid computing, CC is service-oriented; it offers virtualized resources on demand as quantifiable and billable utilities [9,10]. Consider one’s experience with email as an example of CC. The email client, whether it be Yahoo!, Gmail, or Hotmail, manages the hardware and software required to operate one’s personal email account. Emails are not stored on someone’s physical computer; they are accessed via an internet connection, which means they are accessible from any location [11,12].

Figure 1 illustrates how, in a CC environment, various accessories such as mobile phones, personal computers/laptops, and servers are used to share information via frame relays, routers, and switches [13,14]. Any subscriber with an internet connection can access the public cloud, as shown in Figure 2. In the simplest terms, public cloud services are those that are made available to clients via the internet by a third-party service provider. The moniker ‘public’ does not always imply that it is free, even if it is possible to utilize it for free or at a low cost. A public cloud does not imply that a user’s data is publicly accessible; public cloud companies often provide means for their users to manage access to their own data [15].

Figure 3 depicts a private cloud that is created for a specific group or organization and restricts access to that group or organization. A private CC environment shares many characteristics with a public CC environment, including adaptability and a service-oriented architecture. The difference between a private cloud and a public cloud is that, with a private-cloud-based service, data and processes are maintained within the organization, without the network capacity, security risks, and regulatory requirements that are associated with public cloud services [16].

As depicted in Figure 4, a community cloud is a cloud that is shared by two or more enterprises with comparable cloud requirements. A set of organizations with common interests, such as specialized security requirements or a shared mission, will manage and use a community cloud [17].

PaaS, IaaS, and SaaS are the three service types that comprise the CC concept. PaaS is a model in which a third-party provider provides hardware and software tools to customers via the internet. These tools are typically necessary during the application development process. PaaS offers an application’s execution environment, such as the Google App Engine, with point-and-click capabilities that allow non-developers to create their own web applications [18,19]. Second, in IaaS, online services that provide high-level Application Programing Interfaces (APIs) are used to access numerous low-level aspects of the underlying network infrastructure, such as physical computer resources, location, data partitioning, scaling, security, and backup. Thirdly, SaaS enables consumers to access cloud service providers’ applications via a web browser that runs on cloud infrastructure. As a result, end users are not required to download, install, configure, and operate software applications on their personal computing terminals [20,21,22].

Providing rankings involves assigning a value to each option and then arranging them according to that value, with the lowest value representing the best choice. The rank increases as the value decreases. Rankings of cloud services have been gaining popularity over time. However, ranking is subtly different, due to the naming convention and existing cloud infrastructure [23,24]. When there are numerous service providers, selecting a single CC service can be challenging. It is essential to select the optimal ranking methodology and to evaluate all qualitative aspects of the services [25].

In Artificial Intelligence (AI), the creation of a precise prediction model requires the collection of input characteristics that influence cloud-service rankings. However, a significant portion of the required data is inaccessible or incomplete, due to data source retrieval or significant data loss. Therefore, a model with a high degree of accuracy must be created using a limited number of input parameters [26,27]. Various methodologies, including deterministic and probabilistic approaches, have been utilized to estimate the global ranking of cloud services. Few studies have employed LR for estimating cloud service rankings, with multiple LR being the most common technique. This model lacks the ability to comprehend the nonlinearity and complexity of the system’s configuration. To determine the ranking of cloud services, it is preferable to use cutting-edge ML algorithms, which are far more effective and require much less computing time and fewer resources [28,29].

The Machine Learning (ML) tool Waikato Environment for Knowledge Analysis (WEKA) includes an abundance of learning and mining tools. WEKA also makes it easy to test and modify individual learning algorithms to determine which combination of parameters produces the best results [30]. Sequential Minimal Optimization (SMO) is a novel training method for Support Vector Machines (SVM) that divides large problems into as many minor Quadratic Programming (QP) optimization problems as possible [31]. These minor QP issues are resolved analytically, eliminating the requirement for a time-consuming numerical QP optimization in the inner loop. The memory requirements of SMO are proportional to the size of the training set, allowing it to process extremely large training sets, because matrix computation is avoided. SMO scales between linear and quadratic in the training set size for various test problems, whereas the typical SVM technique scales between linear and cubic. SMO can be over a thousand times faster than the typical SVM technique on sparse datasets [32]. For preprocessing, another study utilized wavelet analysis (specifically, the Haar wavelet). It then employed a Support Vector Machine (SVM) via Sequential Minimum Optimization (SMOreg) with Poly kernel function on the updated Quality of Web Service (QWS) dataset to predict cloud-service rankings and model building [33].

2. Cloud Services Based on End-Users’ Feedback