ADsP 51회 대비 족집게 3과목 완전 요약 및 예상문제 대방출 [ 51회 대비 족집게 강의 3강]

Summary

Key Points

• The 50th ADSP exam was challenging, with Subject 3 (Statistical Analysis and Structured Data Mining) featuring a mix of conceptual and data interpretation/calculation questions. 0:23
• Key statistical concepts likely to appear include derived variables, missing values, outliers, sampling methods, data scales, probability calculations, expected values, descriptive statistics, graphs, P-values, hypothesis testing, correlation analysis, and regression analysis. 6:49
• Time series analysis, including stationarity and ARIMA models, and decomposition factors, are also identified as important areas for the exam. 9:49
• Candidates are advised to thoroughly review all covered concepts and practice solving a wide variety of problems, especially those involving calculations and interpretation of results. 10:43
• Topics like multicollinearity in regression and non-parametric tests in statistical analysis are highlighted as potentially appearing due to their recent absence or importance. 44:34
• Understanding the characteristics and assumptions of various models, such as regression analysis error terms and neural network activation functions, is crucial. 51:19
• Statistical Analysis is expected to have around 16 questions, with Structured Data Mining contributing about 14. 66:56
• The video stresses the importance of clearly understanding model characteristics to adapt to potential changes in question notation or format. 68:00
• In Structured Data Mining, emphasis is placed on understanding data mining algorithms (supervised vs. unsupervised), data partitioning, performance evaluation metrics (F1 score, ROC curves), classification models (logistic regression, SVM, decision trees), ensemble methods, artificial neural networks, cluster analysis, and association rule analysis. 69:00
• Calculation problems are increasingly common, particularly for expected values, probabilities, F1 scores, and indices like entropy and Gini impurity. 110:37