The Decision-Making Logic of Big Data Algorithmic Analytics.

Author:Ashander, Laura
  1. Introduction

    Data analytics and machine learning constitutes a class of algorithms that improve or substitute evaluation and decision-making by human beings. (Pagallo, 2018) For machine learning decisions, there are process issues concerning the often-opaque production of results by digital systems, in addition to the operations through which innovative technologies are designed, e.g. the collection of statistics and the conjectures made within algorithms. (Stilgoe, 2018)

  2. Conceptual Framework and Literature Review

    By eliminating human beings from assessments, algorithmic decisions are designed as less discriminatory without the perceived unsoundness, partisanship, or shortcomings of individuals in the decision. (Martin, 2018) The algorithms that process in advance and inspect big data to identify models, directions, and correlations are typically regarded as black-boxed or restricted, but analytics algorithms obtain meaning from, and progressively influence, the society (Carter and Yeo, 2017; Hoffman and Friedman, 2018; Kmecova, 2018; Schinckus, 2018), although, on numerous occasions, that remodeling is semantically blind. (Elragal and Klischewski, 2017)

  3. Methodology and Empirical Analysis

    Building our argument by drawing on data collected from Pew Research Center, we performed analyses and made estimates regarding % of users who say they frequently/sometimes see on social media posts that are overly dramatic or exaggerated/people making accusations or starting arguments without having all the facts/posts that teach you something useful you had not known before/posts that appear to be about one thing but turn out to be about something else and % of users who say, after being directed to view their Facebook "ad preferences" page, that they did not know Facebook maintained this list of their interests and traits/they are not comfortable with Facebook compiling this information/the listings do not very or at all accurately represent them. Data collected from 4,200 respondents are tested against the research model by using structural equation modeling.

  4. Results and Discussion

    The advancement and the application of various knowledge discovery routines and algorithmic investigations necessitate human conjecture and sense making (Pridmore and Hamalainen, 2017). Algorithmic decision-making automatically depends on correlations. Such evidence connections may associate an individual's peculiarities, previous actions, public/private...

To continue reading