Please enable JavaScript.

Coggle requires JavaScript to display documents.

User Story Estimation on the Complexity Decomposition Using Bayesian…

- - - - These categories were formed taking into account the opinion of the authors based on the experience of conducting real software projects
    - - The size could be broken down further by adding two sub nodes, which are: connection functions and data functions.
- - - - First Level Attributes
        These nodes are the ones that collect information for the complexity estimation (Experience with technology, Experience in previous projects, Experience with language, Developer skills, Knowledge of the project theme, Dependence, Connection functions and Data functions).
      - Second Level Attributes
        These attributes help to organize nodes hierarchically.
        (Experience, Technical complexity and Size).
      - User Stories Complexity Attribute/Node
        This shows the final value to measure complexity estimation.
  - - - Survey to determine the importance and value of the attributes within the model
        
        This was to validate the importance and value of the selected attributes within the model.
        
        Applied to 21 professional.
        
        The survey questions were formulated to know the degree of acceptance of the proposed attributes, and questions to collect information regarding the experience of the respondents.
        
        Applied to 19 undergraduate students of computer engineering to know how they differed from the professionals.
        
        Both groups mentioned using Planning Poker.
      - Application of ecuations
        
        Definition 1
        To obtain the non-standardized weight of each first-level attribute: experience with technology, experience with language and experience in previous projects.
        
        Definition 2
        To obtain the normalized weight of each attribute.
        
        Definition 3
        To obtain the weight of the experience attribute and the other second-level attributes.
      - A priori values for first level attributes
        This is which value or state each first level attribute can have. Three scales were used:
        
        Low
        
        Regular
        
        High
      - Construction of Conditional Probability Tables
        A CTP show the probability that an event will occur based on the combination of the attributes and the values of their states. In this step, a CTP is defined for each of their child node of the structure.
- - - - Variables
        It was applied the Spearman correlation test. It was analyzed:
        
        estimationIni: The estimation initial before making the user story.
        
        estimationFin: The estimation final, once the user story is made.
        
        estimationBN: The result obtained by the Bayesian network.
      - Results
        
        The Spearman test for the estimationIni-estimationBN relationship, provided a correlation of 0.659 and a p-value of 0.000.
        
        The estimationFin-estimationBN relationship provided a correlation of 0.799 and p-value of 0.000.
        
        The correlation between the final estimation and the one obtained in the network is greater.
        
        This is a good fact because it indicates that what was obtained by the network is very related to what the student estimates once made user story.