Content uploaded by Feras Al-hawari
Author content
All content in this area was uploaded by Feras Al-hawari on Jun 08, 2019
Content may be subject to copyright.
A preview of the PDF is not available
A Machine Learning Based Help Desk System for IT Service Management
Abstract and Figures
A help desk system that acts as a single point of contact between users and IT staff is introduced in this paper. It utilizes an accurate ticket classification machine learning model to associate a help desk ticket with its correct service from the start and hence minimize ticket resolution time, save human resources, and enhance user satisfaction. The model is generated according to an empirically developed methodology that is comprised of the following steps: training tickets generation, ticket data preprocessing, words stemming, feature vectorization, and machine learning algorithm tuning. Nevertheless, the experimental results showed that including the ticket comments and description in the training data was one of the main factors that enhanced the model prediction accuracy from 53.8% to 81.4%. Furthermore, the system supports an administrator view that facilitates defining offered services, administering user roles, managing tickets and generating management reports. Also, it offers a user view that allows employees to report issues, request services, and exchange information with the IT staff via help desk tickets. Moreover, it supports automatic email notifications amongst collaborators for further action. Yet, it helps in defining business processes with well-defined activities and measuring KPIs to assess the performance of IT staff and processes.
Figures - uploaded by Feras Al-hawari
Author content
All figure content in this area was uploaded by Feras Al-hawari
Content may be subject to copyright.
... Because the CRM system is the base for the classification process, many authors have tackled these in different manners. For instant the author [4] receives the customer information through the help desk application itself, being that the customer is enrolled in it. In this context an artefact was developed that applied vectorization to the term frequency-inverse document frequency (TF-IDF) model and subsequently to a Support Vector Machine (SVM) model for the classification of help desk requests for the German Jordanian University and achieved an accuracy rate of 83%, using description, subject and comments of the requests for classification. ...
This project describes a suggested prototype to carry out the automatic classification of requests from a Port Help Desk. It intents to ascertain if the implementation of this framework is viable for this sector. For this purpose different models were employed, such as SVM, Decision Tree, Random Forest, LSTM, BERT and a SVM hierarchical model. To verify their efficiency these models were evaluated using Precision, Recall and F1-Score metrics. We obtained F1-Scores of 94.36% and 92.48% when classifying the request's category and group respectively. A F1-Score of 93.41% while using a SVM model for category classification when employing a hierarchical classification architecture.
... Advancements in technology, such as AI and ML, have led to the goal of automating this process by predicting the time needed to resolve specific issues based on similar cases in the past. The emergence of ML [4], [5] opens the possibility for automated ticket classification and, thus, enables the prediction of the resolution time needed to solve the cases [6], [7], [8]. Prediction of resolution times in any field is crucial in several domains. ...
The quality of customer service emphasizes support tickets. An excellent support ticket system qualifies businesses to provide clients with the finest level of customer support. This enables enterprises to guarantee the consistency of quality customer service delivered successfully, ensuring all clients have a good experience regardless of the nature of their inquiry or issue. To further achieve a higher efficiency of resource allocation, this is when the prediction of ticket resolution time comes into place. The advancing technologies, including artificial intelligence (AI) and machine learning (ML), can perform predictions on the duration required to tackle specific problems based on past similar data. ML enables the possibility of automatically classifying tickets, making it possible to anticipate the time resolution for cases. This paper explores various ML techniques widely applied in the Resolution Time Prediction system and investigates the performance of three selected ML techniques via the benchmarking dataset obtained from the UCI Machine Learning Repository. Implementing selected techniques will involve creating a graphical user interface and data visualization to provide insight for data analysis. The best technique will be concluded after performing the ML technique evaluation. The evaluation metrics involved in this step include Root Mean Square Error (RMSE) and Root Mean Absolute Error (MAE). The experimental evaluation shows that the best performance among the selected ML techniques is Random Forest (RF).
... Advancements in technology, such as AI and ML, have led to the goal of automating this process by predicting the time needed to resolve specific issues based on similar cases in the past. The emergence of ML [4], [5] opens the possibility for automated ticket classification and, thus, enables the prediction of the resolution time needed to solve the cases [6], [7], [8]. Prediction of resolution times in any field is crucial in several domains. ...
The quality of customer service emphasizes support tickets. An excellent support ticket system qualifies businesses to provide clients with the finest level of customer support. This enables enterprises to guarantee the consistency of quality customer service delivered successfully, ensuring all clients have a good experience regardless of the nature of their inquiry or issue. To further achieve a higher efficiency of resource allocation, this is when the prediction of ticket resolution time comes into place. The advancing technologies, including artificial intelligence (AI) and machine learning (ML), can perform predictions on the duration required to tackle specific problems based on past similar data. ML enables the possibility of automatically classifying tickets, making it possible to anticipate the time resolution for cases. This paper explores various ML techniques widely applied in the Resolution Time Prediction system and investigates the performance of three selected ML techniques via the benchmarking dataset obtained from the UCI Machine Learning Repository. Implementing selected techniques will involve creating a graphical user interface and data visualization to provide insight for data analysis. The best technique will be concluded after performing the ML technique evaluation. The evaluation metrics involved in this step include Root Mean Square Error (RMSE) and Root Mean Absolute Error (MAE). The experimental evaluation shows that the best performance among the selected ML techniques is Random Forest (RF).
... These results showed that the proposed method for campus helpdesk request processing is relatively efficient than the other previous research. [25] 80.3 19.7 Clarin [26] 88.2 11.8 Cheung et al. [27] 84.5 15.5 Alcober et al. [28] 94.5 5.5 Govindan et al. [29] 76.7 23.3 Al-Hawari & Barham [41] 81.4 18.6 Kamal et al. [42] 79.8 20.2 Sharmila et al. [43] 94.6 5.4 Expert-Centric Approach of the proposed method (ECA) 95. 4 4.6 ...
... [5] [6] [7] Several studies discussed the use of machine learning to automate the incident management process, including research conducted by [5] which discusses about determining the priority of incidents using Fuzzy Mamdani. Research [8] [9] [10] also discusses the use of machine learning to determine incident categories. Each study evaluates several supervised learning algorithms with different approaches. ...
Incident Management in ITIL requires an effective process so the incidents do not disrupt business processes for too long. This research aims to automate decision-making in Incident Management process. To perform the automation in decision-making process requires machine learning algorithms. The development of machine learning method in this research will bring significance result such as a new technique of decision-making process in Incident Management, accelerate decision-making process in Incident Management by implementing machine learning to determine the category, group, and priority. By combining supervised and unsupervised machine learning, this research can help to determine the priority of the incident, so IT Operation Teams know which incident should resolve first. By training historical full description, short description, and title, machine learning can classify the new incident. In this research different classification algorithms are used to automate decision making process. Performances of automated decision-making are evaluated with accuracy, precision, recall, and f1-score. Based on the result of various performance metrics, classifier based on K-Nearest Neighbor performed well on predicting Priority, and both category and priority get the best performance with Support Vector Machine.
Over the past ten years, there have been notable advancements in natural language processing (NLP) systems. The goal of the study is to extract relevant key phrases from the Marathi patient major complaints. Our data came from a single, non-profit private hospital located in a remote part of North Maharashtra Region, India. From the hospital's information around 5000 patient visits occurred between 2021 and 2022.A total of 54000 text-free chief complaints were gathered as a result of the initiatives. This paper summarizes a study that focuses on pre-processing strategies that were established in an efficient way to extract pertinent keywords from phrases in accordance with the patients' primary concerns. The methods and findings from this experiment are presented in this paper.
Automation of support ticket classification is crucial to improve customer support performance and shortening resolution time for customer inquiries. This research aims to test the applicability of automated machine learning (AutoML) as a technology to train a machine learning model (ML model) that can classify support tickets. The model evaluation conducted in this research shows that AutoML can be used to train ML models with good classification performance. Moreover, this paper fills a research gap by providing new insights into developing AI solutions without a dedicated professional by utilizing AutoML, which makes this technology more accessible for companies without specialized AI departments and staff.
Automation of support ticket classification is crucial to improve customer support performance and shortening resolution time for customer inquiries. This research aims to test the applicability of automated machine learning (AutoML) as a technology to train a machine learning model (ML model) that can classify support tickets. The model evaluation conducted in this research shows that AutoML can be used to train ML models with good classification performance. Moreover, this paper fills a research gap by providing new insights into developing AI solutions without a dedicated professional by utilizing AutoML, which makes this technology more accessible for companies without specialized AI departments and staff.
e-ticketing is one of the common applications used in technical support in Information Technology (IT) and has been used worldwide in any field of company. The benefits of e-ticketing can reduce the human efforts, increase the sufficiency of system and provides the benefits and efficiency to the customers. Also, e-ticketing ticketing system can enhance worker safety, improving productivity, increasing project efficiency and cause the good impact on the performance of the business in terms of profitability. The main objective of this study is defining the model performance in each important feature by analysing the complex dataset by using logistic regression as a Machine Learning (ML) algorithm. In evaluation the performance of classifier, the dataset is injected to Python programming and split into 90% as training set and 10% for the testing set. From the analysis, the study found that only 3 out of 11 independent features in dataset that are relevant chosen to proceed the ML analysis. From the result, the accuracy for sct_short_description, sct_cmdb_ci, and sct_assignment_group is 41.65%, 48.77% and 96.49%, respectively. It showed that the accuracy’s result for the sct_assignment_group resulted that the model is very good accuracy and indicate that the model is well performing. Meanwhile, the value of F1-score is 96.11% in each feature. This result indicates that the model has a good balance of precision and recall in its binary classification predictions. Hence, the study considers the sct_assignment_group as a best features to proceed the analysis. The future study will consider dealing the combination of complexity features by implementing more analysis on ML such as Support Vector Machines and Naïve Bayes.
The design and development of an information display system are discussed in this paper. This system is used to display dynamic content (e.g., announcements, news, slideshows, and other information) on HD TV screens that are distributed across the university campus. The low cost and small size raspberry pi devices running a Linux-based OS are utilized to open a dynamic web page in web browsers and then transfer the views to the corresponding screens for display. To cut costs further, the management software to update the displayed content is developed in-house and deployed using free open source software technologies such as the Apache server, PHP engine, and MySQL database management system. Furthermore, users may interactively scan a QR code on a screen using a reader running on their smartphones to easily access the offered services.
Incident Management is a key element of supporting any system. For Internet-based applications this function requires integration of staff, process, and tools to manage responses to integrated system events including carrier network outages or performance impacts, hosted server and infrastructure failures, application errors, and mis-configurations. A detailed description of the origin of the Incident Management approach at Wolters Kluwer is discussed, including a summary of the ITIL (Information Technology Infrastructure Library) guidance on which it is based, focusing on the custom process and tools developed to realize an improved approach over time, and the results achieved. Index Terms—Incident Management, Computer Operations, Root Cause Analysis, ITIL, Software Engineering, Process Engineering.
bold xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Background:
This study examined discourse during the problem resolution process in face-to-face technical-support interactions between technical-support providers and users at a helpdesk. Specifically, this study responds to the limited discourse-based research in technical-support interactions, despite the agreement that exploring communication within these interactions can help to improve them.
Literature review:
Research in technical communication has only begun to systematically explore the microlevel (smaller, moment-to-moment) communication in these interactions, though it has provided a well-established understanding of their macrolevel (genre) structure. Further, research has not completely explored how the two participants in the interaction work discursively to resolve technical problems and what strategies appear to promote user satisfaction.
Research questions:
1. What microlevel discourse do technical-support providers employ to resolve technical problems? 2. What microlevel discourse do users employ when collaborating with technical-support providers to resolve technical problems? 3. What microlevel discourse from both technical-support providers and users appears in satisfactory interactions? 4. What exchange patterns between the two participants appear in satisfactory interactions?
Methodology:
Using speech-act discourse analysis, this study examined 17 helpdesk interactions that resolved problems. These interactions occurred at a large, Midwestern US university helpdesk at which 11 instructors sought help with instructional technologies. Using a post-session survey completed by the users, I compared interactions with above- and below-average levels of user satisfaction to determine which microlevel discourse promoted user satisfaction with the help experience.
Results:
Technical-support providers employed signals announcing their thoughts and actions, gave explanations of the technology, and communicated confirmations or denials to user questions. Users employed inquiries about the technology, gave background information, and communicated confirmations or denials to technical-support provider questions. Statistically significant results about both speakers’ discourse indicate that typical instructional strategies (such as explanations) do not necessarily characterize more satisfactory interactions. Instead, alternate forms of instruction (minimal responses or giving background information from personal experience) contribute toward satisfactory outcomes. Also, users’ facility in asking follow-up questions or in giving further background information even during the problem-resolution stage appears to promote satisfaction. Further, a closer examination of the exchanges reveals how both participants interact in these ways.
Conclusions:
This study provides further insight into the typical instructional strategies identified by other researchers. Specifically, although explanations or directions do characterize problem resolution, these discourse strategies do not necessarily characterize distinctly satisfactory problem resolution. As one of the only studies of the functional discourse within technical-support interactions, this study provides researchers and practitioners further insight into how these important interactions work when technical-support providers resolve problems.</italic
The numbers of software products coming out day by day are increasing in the market. Due to the growing demands of the community and to satisfy consumer needs in various domains. In Software Development Lifecycle the steps starts from the collection of requirements, discussion of estimates, allocation of resources, design, development, quality assurance. It so happens that due to high importance for time to market the quality is degraded. The Quality assurance team has to deal with huge number of bugs within a specific timeline and has to triage the bugs and make the bugs which are in common into a group so that the effort is reduced. In this paper first we have a predefined set of bugs across 4 different products. The bugs undergo a series of data mining steps or algorithms namely cleaning, frequency computation, IDFT computation, score computation and then the duplicate bug detection algorithm is applied which groups the bug. The second part of the project is responsible for classification of the bugs in terms of user interface, design, performance and usability. The classification is performed by applying text based naive classifier which finds the probability, contingency and enhanced contingency on the bugs and then assigns a class for the bug.
This paper proposes a simple and efficient approach for text classification and representation learning. Our experiments show that our fast text classifier fastText is often on par with deep learning classifiers in terms of accuracy, and many orders of magnitude faster for training and evaluation. We can train fastText on more than one billion words in less than ten minutes using a standard multicore CPU, and classify half a million sentences among 312K classes in less than a minute.