Research Tools

#⚙️/📜

Zotero

Installation & Env. setting

Research env#5. Zotero

Project managements

조테로 기본 사용 workflow

Indexing ideas within same projects from multiple articles

여러 논문의 정보를 빠르게 인덱싱하는 workflow

refs in core ref.

중심이 되는 저널에 인용된 논문들을 같이 인용할 때의 workflow

CSL

- 괄호 밖에서 구분하기
	<layout delimiter=", ">
		<group prefix="[" suffix="]" delimiter=", ">
        <text variable="citation-number"/>
        <text macro="citation-locator"/>
      </group>
- 괄호 안에서 ; 로 구분하기
       <layout prefix="(" suffix=")" delimiter="; ">
      <group delimiter=", ">
        <group delimiter=" ">
          <text macro="author-short"/>
          <text macro="year-date"/>
        </group>
        <text macro="locator"/>
      </group>
    </layout>

delimiter-precedes-last="always"/>
        <label form="short" prefix=", "/>
        <substitute>
          <names variable="editor"/>
          <text macro="title"/>
        </substitute>
      </names>
    </group>
  </macro>
  <macro name="access">
    <choose>
      <if type="article-newspaper" match="none">
        <choose>
          <if variable="DOI">
            <text value="doi:"/>
            <text variable="DOI"/>
          </if>
          <else-if variable="URL">
            <group delimiter=". ">
              <choose>
                <if type="webpage post post-weblog" match="any">
                  <date variable="issued" prefix="Published " form="text"/>
                </if>
              </choose>

⌨️Hotkeys#Zotero

Obsidian

Plugins

Hotkey

⌨️Hotkeys#Obsidian

Mermaid

About Mermaid | Mermaid
Gantt diagrams | Mermaid

Zettelkasten

Zettelkasten#Obsidian

Overleaf

Draft management

Springer Nature

Table

\begin{table}[htbp] % h: here, t: top, b: bottom, p: page
\centering
\caption{Table Title}
\label{tab2} %본문에서 \ref{tab2}로 연결
\renewcommand{\arraystretch}{1.3} % 행 간격 늘리기
\begin{tabular}{@{}p{1.3cm}p{1.5cm}p{1.5cm}p{1.8cm}p{2.0cm}p{1.8cm}@{}}
\toprule
\textbf{Phase} & \textbf{Active Extension (N=55)} & \textbf{Passive Extension (N=55)} & \textbf{Univariate p-value} & \textbf{Multivariate Estimate (SE)} & \textbf{Multivariate p-value} \\
\midrule
\multirow{2}{*}{APA (mV)} & -0.70 & -0.34 & 0.197 & -- & -- \\
                          & [-1.39; 0.15] & [-0.97; 0.31] & & & \\
\midrule
\multirow{2}{*}{PAPA (mV)} & -1.62 & -0.87 & $<$0.001*** & -0.77 (0.14) & $<$0.001*** \\
                          & [-2.99; -0.94] & [-1.39; -0.43] & & & \\
\midrule
\multirow{2}{*}{RPA (mV)} & 0.30 & 0.39 & 0.986 & -- & -- \\
                          & [-0.65; 0.85] & [-0.17; 0.80] & & & \\
\midrule
\multirow{2}{*}{PRPA (mV)} & 1.27 & 0.81 & 0.011* & -0.46 (0.15) & 0.002** \\
                          & [0.64; 2.38] & [0.40; 1.79] & & & \\
\bottomrule
\end{tabular}
\vspace{1ex}
{\footnotesize
Non-normally distributed continuous variables are summarized with median [interquartile range].\\
SE, standard error; APA, activity phase amplitude; PAPA, post-activity phase amplitude; RPA, return phase amplitude; PRPA, post-return phase amplitude.\\
A generalized linear mixed model (GLMM) was fitted for the multivariable analysis, incorporating the following covariates: age, sex, height, BMI, thigh circumference, and calf circumference.\\
* p-value$<$0.05, ** p-value$<$0.01, *** p-value$<$0.001.
}
\end{table}

Figure

\begin{figure}[h]
\centering
\includegraphics[width=0.9\textwidth]{Figure3.pdf} %첨부한 파일 링크
\caption{\textbf{ Deep learning model structure for EAG classification.} The model takes time series data of size (201, 8, 1) as input. The initial Convolutional Layer uses a 3x3 kernel with 16 filters to extract spatial features and introduces nonlinearity through the ReLU activation function. The following MaxPooling2D layer reduces the spatial dimensions to retain important information. The second Convolutional Layer uses a 3x3 kernel and 32 filters to learn more complex features. This is followed by a Reshape layer that converts the 2D output to a 1D format, leading into Conv1D layers. The first Conv1D layer uses a kernel of size 3 with 32 filters to detect temporal patterns, followed by MaxPooling1D to reduce the time dimension. The second Conv1D layer uses a kernel of size 3 with 64 filters to learn more complex temporal patterns, with another MaxPooling1D layer to further reduce the time dimension. Finally, Fully Connected Layers, including a Flatten layer that converts data into a 1D array, and several Dense layers, allow the model to learn features and perform classification. Dropout layers are employed to prevent overfitting during training. The final output layer, with 8 neurons and a softmax activation function, performs multiclass classification.}
\label{fig3}
\end{figure}

Citations

🔗LINKS TO THIS PAGE
Research env