References don't work before the last section

Question

I am trying to write a paper and after hours of figuring out how to properly work with Biber. Everythink was working well until i moved to the next section. The references disappeared and all that was left was bunch of zeros as seen in the photos. in the next section references are working well but only in that section not the previous one. Here is my code:

\documentclass[conference]{IEEEtran}
\IEEEoverridecommandlockouts
% The preceding line is only needed to identify funding in the first footnote. If that is unneeded, please comment it out.
\usepackage[backend=biber,defernumbers=true,refsection=section]{biblatex}
\usepackage{amsmath,amssymb,amsfonts}
\usepackage{algorithmic}
\usepackage{graphicx}
\usepackage{textcomp}
\usepackage{xcolor}
\def\BibTeX{{\rm B\kern-.05em{\sc i\kern-.025em b}\kern-.08em
    T\kern-.1667em\lower.7ex\hbox{E}\kern-.125emX}}
\usepackage{graphicx}
\usepackage[colorlinks,allcolors=blue]{hyperref}
\addbibresource{biblio.bib}

\begin{document}

\title{Measuring ...}

\author{\IEEEauthorblockN{1\textsuperscript{st} G}
\IEEEauthorblockA{\textit{S} \\
\textit{I}\\
Th \\
.edu.gr}
\and
\IEEEauthorblockN{2\textsuperscript{nd} S}
\IEEEauthorblockA{\textit{School of Science and Technology} \\
\textit{I}\\
Th \\
ihu.edu.gr}}


\maketitle
\begin{abstract}
This document is a model and instructions for latex
This and the IEEEtran.cls file define the components of your paper [title, text, heads, etc.]. *CRITICAL: Do Not Use Symbols, Special Characters, Footnotes, 
or Math in Paper Title or Abstract.

\end{abstract}

\begin{IEEEkeywords}
component, formatting, style, styling, insert
\end{IEEEkeywords}

\section{Introduction}

\subsection{Epileptic...}
The EEG ... intervals and seizure intervals~\cite{alotaiby}. 
Therefore pattern recognition algorithms such as $k$-Nearest Neighbour ($k$-NN) or machine learning Neural Networks are necessary to analyse such data.\cite{liu} 

\begin{figure}
\centering
\includegraphics[width=0.4\textwidth]{fig1.png}
\caption{An EEG signal with containing a seizure~\cite{alotaiby}}
\end{figure}  


\subsection{Data}
Our dataset ...an epileptic seizure~\cite{andrze}. Data available at:~\url{https://archive.ics.uci.edu/ml/datasets/Epileptic+Seizure+Recognition}   

\section{Methodology}
\subsection{Rescaling}
In Machine learning.. [-1,1] with the general form been~\cite{aksoy_resc}:

\begin{align}
x' = \frac{x-min(x)}{max(x)-min(x)}\
\end{align}  

where $x$ is the original value and $x'$ the normalized one.

\subsection{$k$-Nearest Neighbors}
The $k$-nearest neighbors algorithm ($k$-NN) is a non-parametric method used for classification and regression~\cite{altman_knn}. In our problem we have a classification problem. For classification the output of the k-NN is a class membership , in our study is whether we have an epileptic seizure or not. Our object will be classified by the plurality vote of the neighbours with the subject of been epileptic seizure or not, among $k$ nearest neighbours. 

The $k$-NN algorithm is one of the simplest across other machine learning algorithms. In the training phase the algorithm only of storing the class labes and feature vectors of the training sample~\cite{coomans_knn}. Then the user-defined $k$ and a unlabelled vector is classified with the labels that classify the $k$ nearest training sample points. The metric that we used for measuring the distance between the neighbours is the Euclidean distance~\cite{phyu_knn}     


\printbibliography[heading=subbibliography]
\end{document}

Answer 1

The behaviour can be reproduced in the slightly smaller MWE

\documentclass[british]{article}
\usepackage[T1]{fontenc}
\usepackage[utf8]{inputenc}
\usepackage{babel}
\usepackage{csquotes}

\usepackage[backend=biber, style=numeric,
  defernumbers=true, refsection=section]{biblatex}

\addbibresource{biblatex-examples.bib}

\begin{document}
\section{Lorem}
Lorem~\cite{nussbaum}

\section{Ipsum}
Ipsum~\cite{sigfridsson}

\printbibliography
\end{document}

It is caused by the option

refsection=section,

and is exacerbated (luckily, otherwise you might have never realised there is an issue) by defernumbers=true,.

refsection=section, tells biblatex to start a new refsection for each \section (or \section*) command in your document. refsections are designed to produce several completely independent bibliographies in your document. A \printbibliography is always local to the current refsection and will ignore citations from other refsections.

In the example, the \printbibliography belongs to the second \section (\section{Ipsum} in my example, \section{Methodology} in the example from the question) and will only pick up citations from there. The citations from the first section are left hanging without a \printbibliography (note how the nussbaum reference does not appear in any bibliography). Due to the defernumbers=true, the number is only assigned to the citations when they first appear in a \printbibliography. The entries from the first section never appear in any \printbibliography and thus all get the dummy value "[0]".

I assume you want a global bibliography for your document, so the solution is to remove the option

refsection=section,

If you only have a single bibliography list, you can additionally remove the option

defernumbers=true,

---

I should add that IEEEtran is a document class for submissions to IEEE journals and proceedings that comes with its own BibTeX .bst styles. If you want to submit to the IEEE, you should use their BibTeX styles and not biblatex (t is unlikely that the IEEE can and will accept biblatex submissions since they require a different workflow: Biblatex: submitting to a journal). If you don't want to submit to the IEEE, a standard class might be a better choice than IEEEtran.