This is an elective course titled "Fundamentals of AI and Data Analysis" offered by the Department of Finance at Chung Hua University. While data analysis is a common requirement across disciplines, it is rarely structured as a core course in non-technical programs. However, for students in management-related fields, mastering data analysis tools and techniques is essential for interdisciplinary competence.
Big data and artificial intelligence (AI) represent dominant trends in both technology and application. For finance students specializing in financial information systems, the focus lies in applying big data and AI to data analytics. Traditionally, business students have relied on courses like statistics, econometrics, and statistical software tools to interpret data meaningfully.
However, classical statistical methods fall short when applied directly to big data due to scalability issues, as highlighted in critiques like “Statistics Is Not a Tool for Big Data Analysis.” In modern contexts, AI has become foundational to big data processing. For management students, this shift implies the need to acquire AI-based data analysis skills—not just theoretical knowledge.
This course adopts an approach where AI is integrated into data analysis methods, rather than relying on pre-built mathematical models. The AI models taught here are capable of autonomously identifying data patterns and constructing verifiable, interpretable mathematical representations. Unlike traditional AI pipelines that train data against predefined models, this method allows data itself to drive model selection and development, thereby reducing human bias and increasing adaptability.
This is an elective course titled "Fundamentals of AI and Data Analysis" offered by the Department of Finance at Chung Hua University. While data analysis is a common requirement across disciplines, it is rarely structured as a core course in non-technical programs. However, for students in management-related fields, mastering data analysis tools and techniques is essential for interdisciplinary competence.
Big data and artificial intelligence (AI) represent dominant trends in both technology and application. For finance students specializing in financial information systems, the focus lies in applying big data and AI to data analytics. Traditionally, business students have relied on courses like statistics, econometrics, and statistical software tools to interpret data meaningfully.
However, classical statistical methods fall short when applied directly to big data due to scalability issues, as highlighted in critiques like “Statistics Is Not a Tool for Big Data Analysis.” In modern contexts, AI has become foundational to big data processing. For management students, this shift implies the need to acquire AI-based data analysis skills—not just theoretical knowledge.
This course adopts an approach where AI is integrated into data analysis methods, rather than relying on pre-built mathematical models. The AI models taught here are capable of autonomously identifying data patterns and constructing verifiable, interpretable mathematical representations. Unlike traditional AI pipelines that train data against predefined models, this method allows data itself to drive model selection and development, thereby reducing human bias and increasing adaptability.
This is an elective course titled "Fundamentals of AI and Data Analysis" offered by the Department of Finance at Chung Hua University. While data analysis is a common requirement across disciplines, it is rarely structured as a core course in non-technical programs. However, for students in management-related fields, mastering data analysis tools and techniques is essential for interdisciplinary competence.
Big data and artificial intelligence (AI) represent dominant trends in both technology and application. For finance students specializing in financial information systems, the focus lies in applying big data and AI to data analytics. Traditionally, business students have relied on courses like statistics, econometrics, and statistical software tools to interpret data meaningfully.
However, classical statistical methods fall short when applied directly to big data due to scalability issues, as highlighted in critiques like “Statistics Is Not a Tool for Big Data Analysis.” In modern contexts, AI has become foundational to big data processing. For management students, this shift implies the need to acquire AI-based data analysis skills—not just theoretical knowledge.
This course adopts an approach where AI is integrated into data analysis methods, rather than relying on pre-built mathematical models. The AI models taught here are capable of autonomously identifying data patterns and constructing verifiable, interpretable mathematical representations. Unlike traditional AI pipelines that train data against predefined models, this method allows data itself to drive model selection and development, thereby reducing human bias and increasing adaptability.
This is an elective course titled "Fundamentals of AI and Data Analysis" offered by the Department of Finance at Chung Hua University. While data analysis is a common requirement across disciplines, it is rarely structured as a core course in non-technical programs. However, for students in management-related fields, mastering data analysis tools and techniques is essential for interdisciplinary competence.
Big data and artificial intelligence (AI) represent dominant trends in both technology and application. For finance students specializing in financial information systems, the focus lies in applying big data and AI to data analytics. Traditionally, business students have relied on courses like statistics, econometrics, and statistical software tools to interpret data meaningfully.
However, classical statistical methods fall short when applied directly to big data due to scalability issues, as highlighted in critiques like “Statistics Is Not a Tool for Big Data Analysis.” In modern contexts, AI has become foundational to big data processing. For management students, this shift implies the need to acquire AI-based data analysis skills—not just theoretical knowledge.
This course adopts an approach where AI is integrated into data analysis methods, rather than relying on pre-built mathematical models. The AI models taught here are capable of autonomously identifying data patterns and constructing verifiable, interpretable mathematical representations. Unlike traditional AI pipelines that train data against predefined models, this method allows data itself to drive model selection and development, thereby reducing human bias and increasing adaptability.
Big Data Analytics and the Role of AI in Time-Series Regression
Big data analytics refers to the integration of large-scale datasets with diverse analytical methods. The evolution from traditional data analysis to big data analysis—and further into the domain of big data analytics within computer science—marks a shift from method-driven approaches to tool- and technology-driven solutions. However, for professionals in management, the emphasis remains on understanding analytical methods, interpreting results, and applying insights to decision-making.
Therefore, big data courses tailored for management students differ from those offered in computer science or information departments. The focus is placed more on data interpretation and decision relevance than on algorithmic implementation.
As big data technologies converge with artificial intelligence (AI), AI has shown promising applications in text, image, and multimedia classification. Yet, in numerical data analysis—particularly in business contexts—AI has not yielded equally strong outcomes. Since numerical (structured) data remains critical and objectively analyzable, there is growing interest in how AI can be effectively applied to its analysis.
Most business datasets, especially those accumulated over time, are time-series in nature. These carry embedded temporal patterns that can be essential for strategic decisions. However, how AI can incorporate temporal elements into analytical models remains a key challenge.
Regression Analysis in Time-Series Data
The most common method for analyzing time-series data in statistics is regression analysis, typically linear regression (simple or multiple).
Assumptions and Model:
Model: Y = E(Y|X) = β0 + β1 X + ε
Goal: Estimate β0 and β1.
Key assumptions:
Linearity in conditional expectation
Normal distribution of residuals
Homoscedasticity (equal variance)
Independence of observations
Estimation Techniques:
Ordinary Least Squares (OLS): Minimize the sum of squared residuals
Maximum Likelihood Estimation (MLE): Based on normality assumption, maximize the likelihood function
Hidden Assumptions in Practice:
In practice, analysts make implicit decisions:
How many data points (X, Y) to include
Load and process all data at once to generate a single regression line
Problem Identified:
These decisions are typically based on subjective judgment or experience. Analysts decide data partitions manually, and the generated regression line reflects these fixed assumptions.
AI as an Alternative:
What if AI could autonomously determine the optimal number of data points for a regression line based on statistical principles (e.g., minimizing error)? Rather than producing a single regression line from a predefined dataset, AI might generate multiple regression lines, each optimized for different segments of the data. This dynamic segmentation allows for more accurate modeling, especially when the data is heterogeneous or time-dependent.
By replacing the analyst’s manual judgment with AI, the system can adaptively identify where distinct models are needed—based on the same foundational assumptions of regression (like minimal error)—but with greater flexibility and objectivity.
「數學AI建模」軟體引進課程當中，可解決企業面對時間序列的數字型數據所乘載之時間因素，並且還能提供較統計學之迴歸分析更多的結果。
對數據分析、大數據分析、人工智慧有基本的認識
了解數字型的數據可運用統計學迴歸分析幫助找到時間為自變數下的迴歸線(又可稱為趨勢線)。
每週課程以股價為標的，進行數據更新後產生精準趨勢。
長期追蹤的精準趨勢進行前後比對，以及獲得最新趨勢之特徵起訖日期、趨勢方向、持續時間、平均變動情況。
⭐起訖日期會因為新趨勢形成，變成趨勢轉折點。
⭐趨勢持續時間不同於傳統各種分析方法，真實呈現數據規律特徵，所以每條趨勢線的持續時間可能相同也可能不同。
⭐平均變動情況，即趨勢斜率。相較於傳統分析方法更具備趨勢期間內對數據的代表性。
透過做中學的方式，讓學生在每週動手更新數據與操作軟體的過程中，習慣軟體使用與更新數據，並且因為將可自動化操作程序的動作，以教育軟體呈現，讓學生了解自動化流程中的每個程序，知曉所有程序與人工智慧計算、判斷都是可被客觀驗證的。