User Feedback Analysis

• Explicit Feedback

  • Reviews

  • Social Media

• Implicit Feedback

  • Interaction Data

• is the documented opinion of a user adressing any aspect of software or service provided by entity

• contains opinion of user

• mostly free written

• more requirements can be clarified

• study example

  • monitoring case and collect user feedback for 4 months

  • organize workshop with two phases

  • use ontology mediated queries

  • study found more requirements

Data Collection & Preparation —> Feature Engineering —> Model Selection and Training -> Model Evaluation —> Model for Prediction

• General

  • High quality data important

  • Carefully select machine learning performance

• Traditional Machine Learning

  • Few Data & benefits from expert knowledge

  • High Effort in Feature Engineering

• Deep Learning

  • Exceels with much data

  • Not require domain expert knowledge

• Filter

• Allocate & assign

• Compare

• Track over time

• Qualitative analysis of data

  • research method to determine presence of certain themes with given data

• Inductive Content Analysis

  • Bottom-up analysis, involves many passes over data

• Deductive Content Analysis

  • Top-down analysis, high level previous knowledge