Sökning: "interpretability"

Visar resultat 1 - 5 av 120 uppsatser innehållade ordet interpretability.

  1. 1. The Power of Credit Scoring: Evaluating Machine Learning and Traditional Models in Swedish Retail Banking

    Master-uppsats, Göteborgs universitet/Graduate School

    Författare :Emma von der Burg; Saga Strömberg; [2023-06-29]
    Nyckelord :;

    Sammanfattning : In this paper, we investigate and compare different credit scoring models, with special attention paid to machine learning approaches outperforming traditional models. We explore a recently proposed method called the PLTR model, which is a combination of machine learning and traditional logistic regression. LÄS MER

  2. 2. Neural Network-based Anomaly Detection Models and Interpretability Methods for Multivariate Time Series Data

    Master-uppsats, Stockholms universitet/Institutionen för data- och systemvetenskap

    Författare :Deepthy Prasad; Swathi Hampapura Sripada; [2023]
    Nyckelord :multivariate - time series; anomaly detection; neural networks; autoencoders; interpretability; counterfactuals;

    Sammanfattning : Anomaly detection plays a crucial role in various domains, such as transportation, cybersecurity, and industrial monitoring, where the timely identification of unusual patterns or outliers is of utmost importance. Traditional statistical techniques have limitations in handling complex and highdimensional data, which motivates the use of deep learning approaches. LÄS MER

  3. 3. Generating an Interpretable Ranking Model: Exploring the Power of Local Model-Agnostic Interpretability for Ranking Analysis

    Magister-uppsats, Stockholms universitet/Institutionen för data- och systemvetenskap

    Författare :Laura Galera Alfaro; [2023]
    Nyckelord :Explainable Artificial Intelligence; Learning To Rank; Local ModelAgnostic Interpretability; Ranking Generalized Additive Models;

    Sammanfattning : Machine learning has revolutionized recommendation systems by employing ranking models for personalized item suggestions. However, the complexity of learning-to-rank (LTR) models poses challenges in understanding the underlying reasons contributing to the ranking outcomes. LÄS MER

  4. 4. A Dual-Lens Approach to Loss Given Default Estimation: Traditional Methods and Variable Analysis

    Master-uppsats, KTH/Matematik (Avd.)

    Författare :William Jaeckel; Nicolai Versteegh; [2023]
    Nyckelord :Loss given default; estimering; jämförande studie; variabelanalys; kreditförvaltning; utlåning till små och medelstora företag; riskanalys; Loss given default; estimering; jämförande studie; variabelanalys; kreditförvaltning; utlåning till små och medelstora företag; riskanalys;

    Sammanfattning : This report seeks to thoroughly examine different approaches to estimating Loss Given Default through a comparison of traditional estimation methods, as well as a deeper variable analysis on micro, small, and medium-sized companies using primarily regression decision trees. The comparative study concluded that estimating loss given default depends heavily on business-specific factors and data variety. LÄS MER

  5. 5. Towards gradient faithfulness and beyond

    Master-uppsats, Högskolan i Halmstad/Akademin för informationsteknologi

    Författare :Vincenzo Buono; Isak Åkesson; [2023]
    Nyckelord :XAI; Visual Explanations; CAM; Grad-CAM; Expected Grad-CAM; Hyper Expected Grad; Class Activation Maps; Explainable AI; Faithfulness; Neural Network interpretability; Hyper Resolution CAM; Super Resolution CAM; Natural Encoding;

    Sammanfattning : The riveting interplay of industrialization, informalization, and exponential technological growth of recent years has shifted the attention from classical machine learning techniques to more sophisticated deep learning approaches; yet its intrinsic black-box nature has been impeding its widespread adoption in transparency-critical operations. In this rapidly evolving landscape, where the symbiotic relationship between research and practical applications has never been more interwoven, the contribution of this paper is twofold: advancing gradient faithfulness of CAM methods and exploring new frontiers beyond it. LÄS MER