👨‍🏫جلسه ۱: آموزش انتخاب مقاله پایه و تعیین موضوع پروپوزال

Main ideas / lessons

• Purpose of the training session: Learn how to write a full research proposal by moving from:
  • choosing a research topic → to shaping the proposal → and (later) preparing presentation materials for defense.
• Core problem students face: Many students struggle to:
  • pick a topic,
  • pressure professors too much,
  • or rely on outside researchers to identify a “final” topic.
• What a proposal should do: A proposal should either:
  • introduce a new idea, or
  • improve existing methods through a research process.
• How to narrow a broad field into an approvable topic: The session uses a step-by-step “funnel”:
  1. choose a general area
  2. choose a more specific focus
  3. choose an application domain/system
  4. narrow to a particular disease/task
  5. refine into a concrete topic supported by recent, valid articles

---

Methodology / step-by-step topic selection process (detailed)

Step 1: Identify your general field and interests

• Determine which part(s) of your major you care about most.
• Example (computer science): cloud computing, data mining, wireless sensor networks, SDN, IoT.

Step 2: Use Google to understand the focus and find challenges

• Search broad terms like: “data mining in computer science” and related “data mining problem/challenges”.
• Identify what the area is and what major issues exist.

Step 3: Convert the general topic into an “application”

• Ask: Where will this research be applied?
• Examples:
  • data mining → customer relationship management (CRM)
  • data mining → marketing
  • data mining → medicine

Step 4: Make it medically specific (application → task)

• Narrow further: medicine → a specific medical domain/problem.
• Examples mentioned:
  • cancer diagnosis
  • Persian examples like:
    • “Techniques for diagnosing leukemia using data mining”
    • “Diagnosing breast cancer using data mining”
    • “Diagnosing stomach cancer”
    • “Diagnosing diabetes” (and other diseases)
  • Further specialization examples:
    • diagnosing breast cancer, stomach cancer, leukemia
    • diagnosing MS using MRI images

Step 5: Produce multiple candidate topics (consult several proposals)

• Prepare a small set (e.g., 3–4 topics) such as:
  • “data mining-based approach to diagnosing cancer”
  • “data mining-based approach to diagnosing stomach cancer”
  • “data mining-based approach to diagnosing breast cancer”
• Goal: present several options to the professor.

Step 6: Match the professor’s expectations with article requirements

• The professor may require 1–2 recent supporting articles per candidate topic (examples referenced: 2018/2019).
• Use reputable sources/publishers such as:
  • Elsevier
  • Springer
  • Wiley
  • (also mentioned: “iTel” / unclear site name due to subtitle errors)

Step 7: Convert the topic into English (and search effectively)

• Emphasis: many professors want English articles.
• The session mentions using a translation site (example shown: LG.com).
• Convert topic terms into English for searching.

Step 8: Use Google Scholar to find downloadable articles

• The session references a prior Ryan Thesis training video:
  • “Learning How to Search for Persian and Latin Articles in Google Scholar” (~17 minutes; price mentioned in subtitles).
• Workflow:
  • copy the Persian topic (or translate to English)
  • paste into Google Scholar
  • open relevant results and use available download options

Step 9: Extract ideas from abstracts and diagrams to refine your topic

• From each paper’s abstract, identify:
  • dataset source
  • reported accuracy/metrics
  • the method used
• Use the paper structure to guide what to extract:
  • input data
  • preprocessing
  • classification/modeling step
  • evaluation criteria (accuracy, precision/recall, detection rate)
  • discussion/conclusion
• Then propose a newer or improved method relative to what the paper used (e.g., replacing an older technique with a current one).

Step 10: Search and filter using major databases (example: ScienceDirect)

• Example workflow:
  • go to ScienceDirect (Elsevier)
  • search keywords (e.g., cancer + data mining)
  • filter for recent years (e.g., “bring me articles from 2019”)

Step 11: Mentions a “free download” strategy

• The session states some articles require payment (example shown: $35.95).
• It then claims there is a way to download without paying via a “freeme” type service.
• Subtitles describe a flow:
  • copy the article link
  • paste it into a free-download site
  • enter random security characters
  • download appears to be provided

    Note: The exact method and site details are unclear due to subtitle errors.

Step 12: Build the final topic using method upgrades

• After choosing a base topic, refine it by adding a specific ML approach.
• Example progression:
  • “data mining-based approach to breast cancer diagnosis”
  • variants such as:
    • reinforcement learning
    • hybrid machine learning
    • deep learning
  • additional examples:
    • genetic algorithm
    • ant colony or combinations of algorithms
• Also adjust the year range:
  • if 2019 doesn’t work, try 2018/2017, etc.
  • aim for relevance + novelty.

Step 13: Present the chosen topic(s) to the professor and finalize

• Provide:
  • your refined topic title
  • supporting recent articles
  • a justification for novelty (new method / updated approach)
• The instructor claims approval likelihood is higher if the topic is new and from reputable sources.

Step 14: Insert the approved topic into the proposal template

• The proposal form/template is described as a universal template usable across disciplines (referred to in subtitles as from “University of Research Sciences”).
• After approval (example: hybrid machine learning approach):
  • copy the topic/title into the Persian proposal
  • proceed with the research more confidently

---

Key concepts emphasized

• Narrowing strategy: broad field → focus → application domain → specific disease/task → concrete method-based topic.
• Novelty via method improvement: replace/upgrade the technique used in prior work.
• Use of academic evidence: supporting articles, recent publications, and reputable journals.
• English readiness: translate the topic to access a wider literature base.
• Proposal development pipeline: topic selection is only the first step; later parts include:
  • problem statement
  • literature review
  • objectives/questions/hypotheses
  • methodology
  • keywords
  • referencing

---

Speakers / sources featured (as stated or implied)

• Speaker/Presenter: Instructor for the Ryan Thesis official website (no personal name given in subtitles).
• Organization/Sources referenced:
  • Ryan Thesis
  • Google (initial searching)
  • Google Scholar
  • Elsevier (including ScienceDirect)
  • Springer
  • Wiley
  • LG.com (translation site mentioned)
  • Freeme (free-download service mentioned; URL unclear)
  • “Translate site” (generic mention of translation)
  • University of Research Sciences (proposal template/form reference)