##plugins.themes.bootstrap3.article.main##

This paper explores the synergy between Kaizen methodologies and Machine Learning, with the goal of improving the efficiency of categorizing areas of improvement. The implementation of Machine Learning algorithms, utilizing the ML.NET tool, has proven to be highly advantageous. These algorithms automate the categorization of Kaizen proposals based on structured data and predefined labels. This automation significantly reduces manual workload, expedites the classification process, and ultimately accelerates decision-making within the realm of continuous improvement. Furthermore, the application of machine learning has improved the categorization accuracy, reduced the risk of human errors, and ensured a more precise assignment to the relevant categories. This, in turn, enhances the quality of the collected information, facilitating well-informed decision-making. Additionally, the automated categorization has streamlined the process for users submitting Kaizen proposals, eliminating the need for manual category selection. This reduces cognitive load and promotes active engagement in the continuous improvement process. Furthermore, the system has introduced a category-based incentive structure, where users accrue points corresponding to the category assigned to their Kaizen proposals. These points can be exchanged for items provided by the company, acting as an additional incentive for active employee participation in the continuous improvement process. In conclusion, the integration of Machine Learning with the ML.NET tool has become a valuable tool for optimizing the management of improvement proposals, resulting in a marked enhancement in efficiency, accuracy, and user engagement in the categorization process.

Downloads

Download data is not yet available.

Introduction

The philosophy of continuous improvement, known as Kaizen, has become an essential approach for process optimization in organizations worldwide. The effectiveness of Kaizen largely depends on the efficient identification and categorization of improvement areas within an organization. Traditionally, this task has fallen upon teams of experts in the field, often involving a high degree of subjectivity. However, recent advances in the field of Machine Learning have opened new opportunities to expedite and enhance this process through automation. Several previous studies have explored the potential of Machine Learning in enterprise areas, emphasizing its ability to discern patterns in proposals and assign appropriate categories. In this context, we will delve into the burgeoning influence of Machine Learning and its role in the transformation of various aspects of our society and the business world.

Machine Learning has become one of the mainstays of Information Technology and with that, a rather central, albeit usually hidden, part of our life. With the ever-increasing amounts of data becoming available there is a good reason to believe that smart data analysis will become even more pervasive as a necessary ingredient for technological progress [1].

Machine Learning is a field that has revolutionized artificial intelligence, changing the way we tackle complex problems and enabling decision-making across many industries. The core of machine learning is teaching computers to learn and evolve automatically based on prior knowledge and experience. By allowing machines to recognize patterns, make decisions from data and complete tasks without human intervention, machine learning has led to significant advances in fields such as medicine, economics, robotics, and data analytics. This discipline has become a key tool for solving complex problems and making informed decisions in the digital age.

The primary goal of a Machine Learning implementation is to develop a general-purpose algorithm that solves a practical and focused problem. Machine Learning algorithms are also classified under these learning problems [2]. Depending on the specific learning objective, the field provides a wide array of Machine Learning algorithms, each with numerous variations and specifications, including regressions models, instance-based algorithms, decision trees, Bayesian methods, and ANNs (Artificial Neural Networks) [3].

Fig. 1 showcases different branches within the field of Machine Learning. These subfields are one of the ways Machine Learning algorithms are classified [2].

Fig. 1. Subfields of machine learning. Gollapudi [2], p. 21.

According to Gerón [4], Machine Learning (ML) comprises four primary categories, specifically: (1) supervised learning; (2) unsupervised learning; (3) semi- supervised learning; and (4) reinforcement learning. In supervised learning, the training set used to feed the algorithm includes the desired solutions, called labels. On the other hand, in unsupervised learning the training data are unlabeled, and the algorithm must be learned without prior assistance [5].

Supervised Machine Learning algorithms are those algorithms which need external assistance. The input dataset is divided into train and test datasets. The training dataset has an output variable which needs to be predicted or classified. All algorithms learn some kind of patterns from the training dataset and apply them to the test dataset for prediction or classification [6]. Supervised learning is the most common technique in classification problems since the goal is often to get the machine to learn a classification system that we’ve created [7].

Machine Learning relies on algorithms to analyze huge datasets. Machine Learning is only part of what a system requires to become an AI. The Machine Learning portion of the picture enables an AI to perform these tasks: Adapt to new circumstances that the original developer didn’t envision Detect patterns in all sorts of data sources, create new behaviors based on the recognized patterns and make decisions based on the success of failure of these behaviors [8].

This article delves into the intersection of Kaizen and Machine Learning within the broader domain of Machine Learning. It investigates how the integration of Machine Learning techniques can amplify the efficacy of categorizing areas for improvement. The utilization of Machine Learning algorithms facilitates immediate data analysis and the revelation of concealed patterns within datasets, leading to decision-making that is not only more agile but also better informed in the pursuit of continuous improvement. As organizations aim to optimize their operations, the proficient adoption of these methodologies can deliver substantial advantages concerning competitiveness and overall business performance.

Related Works

Many research papers and articles offer a comprehensive insight into various methods and algorithms employed in the field of Machine Learning. The author in [7], summarizes the fundamental aspects of a couple of supervised methods. The main goal and contribution of this review article are to present an overview of Machine Learning and provide Machine Learning techniques. This paper provides an overview of a couple of supervised learning algorithms. There is a brief explanation of the Machine Learning process.

A recent study by Mahes [6] indicates that the primary advantage of using Machine Learning is that once an algorithm learns what to do with the data, it can perform its tasks automatically. In this article, a brief review, and a future perspective of the extensive applications of Machine Learning algorithms are conducted, concluding that if you have lesser amount of data and clearly labelled data for training, opt for Supervised Learning. Especially in tasks related to high-dimensional data such as classification, regression, and clustering, Machine Learning shows good applicability. By learning from previous computations and extracting regularities from massive databases, can help to produce reliable and repeatable decisions. For this reason, Machine Learning algorithms have been successfully applied in many areas, such as fraud detection, credit scoring, next-best offer analysis, speech and image recognition, and Natural Language Processing (NLP) [3].

In their article’s conclusion, Osisanwo et al. [1] describe various techniques for supervised Machine Learning classification, compare multiple supervised learning algorithms, and determine the most efficient classification algorithm based on the dataset, the number of instances, and features, using the Diabetes dataset for classification, which consists of 786 instances with eight attributes. Machine Learning algorithms are organized into a taxonomy based on the desired outcome of the algorithm. Supervised learning generates a function that maps inputs to desired outputs.

A supervised scenario is characterized by the concept of a teacher or supervisor, whose main task is to provide the agent with a precise measure of its error (directly comparable with output values). Starting from this information, the agent can correct its parameters to reduce the magnitude of a global loss function [9].

Mohamed in 2017 [10] conducted a comparative study of four well-known supervised machine learning techniques: decision tree, nearest neighbor, artificial neural network, and support vector machine. At the end of the study, a practical application was performed to compare their performance. The study found that different classification techniques were used in various areas with different datasets, each having its own strengths and weaknesses. It’s challenging to find a single classifier that can consistently classify all datasets with equal accuracy.

Furthermore, Sharpe et al. [11] in their research elucidate the use of Machine Learning for engineering design exploration and optimization problems by investigating the performance of popular classification algorithms on a variety of example engineering optimization problems. The results are synthesized into a set of observations to provide engineers with intuition for applying these techniques to their own problems in the future, as well as recommendations based on problem type to aid engineers in algorithm selection and utilization.

Tran and Thai-Nghe [12] proposed a method for automated classification of articles submitted online. The system extracts the author of the article and classifies it. They used natural language processing to preprocess the data and then Machine Learning techniques to classify the articles. With the proposed model, it is feasible to extract information and automatically classify items submitted to a classification system. The results showed that the combination of natural language processing and Machine Learning algorithm (e.g., SVM) is effective to develop an automatic item classification system in general.

On the other hand, Imran et al. [13] in their research work used news sources from Pakistan in two languages (English and Urdu) and using Machine Learning, the category of news title and category of news description could be predicted. Different Machine Learning and feature extraction algorithms were used, finally, they built a model using a Machine Learning algorithm with 89% accuracy with logistic regression.

A recent study by Prabhanjan et al. [14] used various classifiers on different types of labeled documents and assessed the accuracy of the classifiers. They employed an Artificial Neural Network (ANN) method that utilizes a Backpropagation Network (BPN), along with several additional techniques, to create an autonomous policy for the procedure of labeled and supervised text categorization. In the work of Rucco et al. [15] utilized an available standard mechanism to analyze the performance of categorization using labeled documents. The investigative examination of real data revealed that the mechanism works well in terms of categorization accuracy.

In 2018, researchers [16] showed a novel approach to classify Chinese news text classification. They used different Machine Learning algorithms such as KNN, Naïve Bayes and SVM. The study showed that SVM with TF-IDF has the best accuracy of 95.7% for small data sets compared to the Naïve Bayes algorithm, which is easy to implement. KNN is the best one of that data set which has more overlapping categories.

In a wide array of fields and disciplines, supervised Machine Learning algorithms have proven to be an invaluable tool. Regarding the healthcare industry Princy et al. [17] researched, a cardiovascular dataset analyzed using various advanced supervised Machine Learning algorithms tailored for disease prediction. The findings demonstrate that the Decision Tree classification model outperformed other methods, such as Naive Bayes, Logistic Regression, Random Forest, SVM, and KNN, in predicting cardiovascular diseases. The Decision Tree model yielded the highest accuracy at 73%. This method could prove beneficial for healthcare professionals in proactively predicting heart diseases and delivering timely treatment.

Furthermore, Ahmad et al. [18] conducted a study using a Decision Tree (DT), an Artificial Neural Network (ANN), bagging, and Gradient Boosting (GB) to forecast the compressive strength of concrete at high temperatures based on 207 data points. Using the ensemble Machine Learning algorithm was shown to improve the performance level of the model.

Several research papers and articles have extensively expounded upon a wide array of methodologies and algorithms employed within the domain of Machine Learning. The outcomes and discoveries of these investigations have effectively showcased the adaptability and efficacy of Machine Learning algorithms in various contexts, affording a profound understanding of their performance across diverse scenarios. The selection of Machine Learning algorithms largely depends on the characteristics of the data and the specific requirements of each application. This highlights the importance of evaluating performance from multiple perspectives. The successful application of these algorithms in areas such as healthcare, engineering, text classification, and the prediction of physical properties underscores their relevance across various academic fields.

Proposed Method

This study presents a novel approach to effectively categorize areas for improvement delineated within Kaizen proposals. This is accomplished through the application of Machine Learning as a means to refine the classification process. Once the Kaizen proposals are submitted to the proposed system, a detailed description of the Kaizen, the Before, and the After are extracted. Additionally, it is necessary to identify the classification, as shown Fig. 2.

Fig. 2. Diagram of the proposed system.

ML.NET is a free cross-platform and open-source .NET framework designed to build and train Machine Learning models and host them within .NET applications [19]. Furthermore, it provides a developer-friendly API and abstractions that make the construction and training of machine-learning models easier without demanding extensive expertise in the field. Supervised learning was employed to train a model using the ML.Net tool. This tool enables working within a scenario to classify data, using tabular information with predefined labels for prediction.

Fig. 3 depicts the new Kaizen submission form. Within this window, the user provides a brief description of the addressed problem; in the “Before” section, he explains the situation before the implementation of the change or improvement; in the “After” section, he mentions the changes or improvements made and the benefits obtained. “Area of Implementation” is the specific area where the Kaizen was applied, and in “Select Category,” the user selects the primary category to which the Kaizen belongs. These categories earn accumulative points for the user, which can be exchanged for items that the company offers as an incentive for participation.

Fig. 3. Designed user interface.

This system was developed using Angular 15 as the front end and is hosted on the company’s intranet. Users authenticate themselves with their username and password. The system communicates with the back-end system through a Web API. The back-end was developed using C# (.Net 6.0) and the database used is MS SQL Server 2019. To address the issue of category assignment, a multidisciplinary team was formed, which reviewed and approved many Kaizen records in the database. With this information, a Machine Learning model was trained using the ML.NET tool for this specific purpose. Next, we have the “Add Data” section, in which the database to be used for training the Machine Learning model is selected. In this section, it is necessary to specify the database, table, and target column for prediction, as can be seen Fig. 4.

Fig. 4. Add data section.

“Train” section, in which ML.NET trains the model for a specified duration. Once completed, the system displays the model with the most accurate results from the training, which will be used during the project’s implementation as shown in Fig. 5.

Fig. 5. Train section.

The system automatically generates the corresponding code to predict the appropriate category based on the entered description, Before, and After. Once this was completed, modifications were made to the back-end system to determine the corresponding category upon receiving a new Kaizen.

Fig. 6 displays the modified New Kaizen window. It is no longer necessary to define the category, as it will be selected at the time of submission to the Continuous Improvement department, thanks to the implemented changes.

Fig. 6. Screen of the modified new Kaizen.

Results

The utilization of Machine Learning techniques through ML.NET has substantiated significant efficiency gains. The automated classification of Kaizen from structured data and predefined labels alleviates manual workloads and expedites the categorization process. This results in expedited responses to improvement proposals and a reduction in decision-making latency. The application of machine learning has engendered a heightened level of precision in the categorization of Kaizen. By employing a well-trained model for category prediction, the potential for human errors is minimized, ensuring the accurate assignment of Kaizen to its appropriate categories. This, in turn, augments the quality of gathered data and streamlines well-informed decision-making.

The automation of the categorization process has simplified the Kaizen submission procedure for users. The necessity for manual category selection has been obviated, diminishing cognitive burdens, and decreasing the probability of errors. This not only enhances the overall user experience but also stimulates increased engagement in the continuous improvement process.

Furthermore, the automated categorization system has introduced a category-driven incentive framework. Users have the opportunity to amass points based on the categories associated with their Kaizen submissions, which can subsequently be exchanged for items provided by the organization. This serves as an added stimulus for active employee participation in the continuous improvement process.

Conclusion

This research accentuates the profound capacity of Machine Learning within the sphere of perpetual business enhancement. The increased accuracy and efficiency in the classification of Kaizen not only streamline the decision-making process and augment the caliber of accumulated data but also cultivate an environment that is favorable to the development of a strong culture of ongoing improvement. The advantages identified extend beyond operational aspects and hold the potential to positively influence the company’s competitiveness and overall effectiveness. The effective deployment of this solution serves as an illustration of how technology can propel continuous improvement within business contexts.

Future Works

As the world moves into the digital age, the integrity and originality of content will become key factors in all areas of education, science, and innovation. Plagiarism detection has been an ongoing problem. Machine Learning, a branch of artificial intelligence based on the ability of computers to learn from data, has emerged as a powerful tool in a wide variety of applications.

In the field of plagiarism detection, this technology could revolutionize the way we identify the similarity of documents and texts by providing a more sophisticated and automated approach. Applying Machine Learning to plagiarism detection offers many potential advantages. By using Machine Learning algorithms, identification solutions can become more accurate and efficient. In addition, they allow analysis of similarities not only in terms of text but also in terms of structure and style, which could improve the detection of plagiarism in more subtle cases. The future of plagiarism detection based on Machine Learning promises a more sophisticated and effective approach to solving this problem in the digital world.

References

  1. Osisanwo F, Akinsola J, Hinmikaiye J, Olakanmi O, Akinjobi J. Supervised machine learning algorithms: classification and comparison. Int J Comput Trends Technol (IJCTT). 2017;48(3):128–38.
     Google Scholar
  2. Gollapudi S. Practical Machine Learning. Birmingham: Packt Publishing; 2016, pp. 4–14.
     Google Scholar
  3. Janiesch C, Zschech P, Heinrich K. Machine learning and deep learning. Electron Markets. 2021;31:685–95. doi: 10.1007/s12525-021-00475-2.
     Google Scholar
  4. Géron A. Hands-on Machine Learning with Scikit-Learn, Keras, and TensorFlow. Sevastopol, CA: O’Reilly Media; 2019.
     Google Scholar
  5. Dias Canedo E, Cordeiro Mendes B. Software requirements classification using machine learning algorithms. Entropy. 2020;22(9):1057. doi: 10.3390/e22091057.
     Google Scholar
  6. Mahes B. Machine learning algorithms—A review. Int J Sci Res (IJSR). 2020;3(1):381–6. doi: 10.21275/ART20203995.
     Google Scholar
  7. Nasteski V. An overview of the supervised machine learning methods. Horizons. b. 2017;4:51–62. doi: 10.20544/HORIZON S.B.04.1.17.P05.
     Google Scholar
  8. Mueller J, Massaron L. Machine Learning for Dummies. 2nd ed. New Jersey: John Wiley & Sons; 2021, pp. 11–7.
     Google Scholar
  9. Bonaccorso G. Machine Learning Algorithms. Birmingham: Packt Publishing; 2017, pp. 9–14.
     Google Scholar
  10. Mohamed A. Comparative study of four supervised machine learning techniques for classification. Int J Appl Sci Technol. 2017;7(2):5–18.
     Google Scholar
  11. Sharpe C, Wiest T, Wang P, Seepersad CC. A comparative evaluation of supervised machine learning classification techniques for engineering design applications. J Mech Des. 2019;141(12):121404. doi: 10.1115/1.4044524.
     Google Scholar
  12. Tran D, Thai-Nghe N. Article classification using natural language processing and machine learning. 2019 International Conference on Advanced Computing Applications (ACOMP), pp. 78–84, 2019. doi: 10.1109/ACOMP.2019.00019.
     Google Scholar
  13. Imran M, Abubakar M, Hussain S, Naveed H, Munawwar J. Classification of news articles using supervised machine learning approach. Pak J Eng Technol (PakJET). 2020;3:26–30. doi: 10.51846/vol3iss03pp26-30.
     Google Scholar
  14. Prabhanjan B, Ghate S, Harshavardhan A, Jhansi G, Sudheer K, Sudarshan E. Text categorization performance examination using machine learning algorithms. IOP Conf Ser Mater Sci Eng. 2020;981:22044. doi: 10.1088/1757-899X/981/2/022044.
     Google Scholar
  15. Rucco M, Giannini F, Lupinetti K, Monti M. A methodology for part classification with supervised machine learning. AI EDAM. 2019;33(1):100–13. doi: 10.1017/S0890060418000197.
     Google Scholar
  16. Miao F, Zhang P, Jin L, Wu H. Chinese news text classification based on machine learning algorithm. 2018 10th International Conference on Intelligent Human-Machine Systems and Cybernetics (IHMSC), 2018 Aug 48–51. doi: 10.1109/IHMSC.2018.10117.
     Google Scholar
  17. Princy RJP, Parthasarathy S, Hency Jose PS, Raj Lakshminarayanan A, Jeganathan S. Prediction of cardiac disease using supervised machine learning algorithms. 2020 4th International Conference on Intelligent Computing and Control Systems (ICI-CCS), pp. 570–75, 2020. doi: 10.1109/ICICCS48265.2020.9121169.
     Google Scholar
  18. Ahmad A, Ostrowski KA, Ma´ slak M, Farooq F, Mehmood I, Nafees A. Comparative study of supervised machine learning algorithms for predicting the compressive strength of concrete at high temperature. Materials. 2021;14(4222). doi: 10.3390/ma14154222.
     Google Scholar
  19. Esposito D, Esposito F. Programming ML.NET. Pearson Education; 2022.
     Google Scholar