LG Aimers 4th Hackathon

LG_Aimers_4th_Hackathon

👪 Members

👑	1	2	3	4
서상혁	김도연	김다운	신동혁	서준혁

---

🏆 Execution Results

• LG Aimers 4th Online Hackathon
  • Public Rank: 40th (F1 : 0.7623)
  • Private Rank: 102rd (F1 : 0.7482)
  • Competitors: Almost 800 teams

---

📽️ Project Intro

Subject	Introduce a strategic marketing method by selecting B2B customers who can be converted into sales customers.
Processing	Check what each column of the data represents and verify the distribution, missing values, and presence of outliers in each column. After conducting data-specific preprocessing and imputing outliers, review the basic statistics, formulate analysis hypotheses for each group based on progression, and validate them using appropriate methods. Create the optimal model through machine learning modeling.
Development Environment	Tool: Jupyter Notebook, VS Code
Communication Environment	Notion: For role assignment in the project, brainstorming ideas, and recording meeting notes related to the project. SLACK, Zoom, Discord, Offline: For real-time remote meetings and in-person meetings.

---

⚙️ Architecture

Category	Content
Model	(Optuna +) Boosting Model, XGBoost, LightGBM, CatBoost, AutoML
Data	Data: Data provided by LG Aimers (Confidential)
Model Evaluation	Calculate the F1-score and select the model with the highest value.
Model Application Method	For each model, compute the final public score and then apply Ensemble or Stacking using the model with the highest F1-score.

---

📆 Project Procedure

You can check the detailed progress on our team's Notion.

https://sixth-drum-9ac.notion.site/LG-Aimers-4-fae749aa84c047699e35cb406dd4e6c2?pvs=4

---