Multiclass Classification of Dermatology Disorders Using Logistic Regression: A Comparative Study of One vs One and One vs Rest Approaches
Abstract
This research paper addresses the challenging task of multiclass classification within the realm of dermatology, employing the versatile Logistic Regression algorithm. The dataset under investigation comprises six distinct classes representing various dermatological disorders. The primary objectives of this study are twofold: to evaluate the performance of Logistic Regression when implemented with the One vs One and One vs Rest strategies, and to assess their effectiveness in classifying dermatological conditions accurately. Results indicate that both strategies exhibit remarkable classification accuracy, precision, and recall rates, underscoring their potential in dermatological diagnosis.
Keywords
Download Options
Introduction
Skin problems are a serious overall general medical problem that influences an enormous number of people. As of late, with the quick headway of innovation and the utilization of various information mining draws near, treatment of skin prescient arrangement has truly become exceptionally prescient as well as exact. The main component of the human body is the skin. The skin saves the body from UV beams, diseases, wounds, temperature, and harming radiation, as well as supporting nutrient D3 [1]. Since the skin is so fundamental in overseeing center temperature and shielding the body from skin problems, it's essential to keep it sound. Skin issues might seem innocuous, yet they can be risky on the off chance that not treated as expected. Numerous illnesses have early side effects, however the majority of them are indistinguishable, making it hard to analyze the condition at an underlying point. Skin problems cause physical as well as mental issues, especially in people whose appearances have been scarred or distorted. Skin can be impacted by a scope of outside and inward factors. Counterfeit skin injury, extreme substance causes, difficulty sicknesses, an individual's invulnerability, and hereditary oddities are a portion of the elements that impact skin issues. Skin illnesses affect individuals' lives including great [2]. Dermatological ailments are the most difficult subfields of science to fix due to the confusions in treating side effects and how side effects modify in different circumstances [6]. Skin sicknesses are regular among numerous ailments, and on the off chance that these strategies are not good for that type of skin condition, it will cause adverse consequences. Individuals are habitually contaminated by skin diseases, which should be treated straightaway. Thus, the kind of AI approaches prepared to do effectively separating skin condition arrangement is fundamental. Up to this point, nobody AI approach has outflanked the others with regards to skin sickness expectation.
Conclusion
This study demonstrates the efficacy of Logistic Regression in multiclass classification for dermatological disorder diagnosis, employing both One vs One and One vs Rest strategies. These models exhibit high accuracy, precision, and recall rates, underlining their potential as valuable tools in the field of dermatology. Further research and clinical validation may enhance their applicability and impact in real-world healthcare scenarios.