Final project structure:
-
Introduction: 1.1. General background on the business or the field from which the problem came. 1.2. Description of the problem from a business point of view. 1.3. Description of the data mining problem.
-
The business problem: 2.1. Description of the problem in business terms. 2.2. Description of the current situation in the business. 2.3. What business goals / why do we need the artificial intelligence solution?
-
Artificial intelligence: 3.1. Description of the problem in terms of data mining/artificial intelligence. 3.2. Determining the objectives of the data mining/outputs of the system.
-
The data: 4.1. A detailed description of the required data includes the format, expected sizes, and required fields. 4.2. A detailed description of the data collection methods. 4.3. Assessment of data quality: missing data values, accuracy, and reliability. 4.4. Building data if necessary: attributes derived from other attributes, changing the values of existing attributes, completing missing values. 4.5. If there are separate tables, this should be considered.
-
Models: 5.1. Review the models/algorithms that can solve the problem (up to 3 models). 5.2. Description of the selected algorithm - a theoretical description as it exists in the literature; the reason for choosing this should be addressed. 5.3. Description of the selected algorithm in the context of the current problem (some algorithms require special adjustments such as Working with different types of data (for example: 5.3.1. Determining the parameters of the model - if necessary. 5.3.2. Determining/description of a distance function - if necessary. 5.4. Describing the output of the system.
-
Implementation of the algorithm 6.1. Implementation of the algorithm on real/synthetic data.
It is allowed to use existing libraries.
The choice of tool you will use to visualize the algorithm is up to you.
It is allowed to use the Internet and existing open-source code - the source must be cited.
7. Evaluation of the model:
7.1. The percentage of error.
7.2. Weighted assessment - f-score.
7.3. Explanations of reasons for the success and failure of the model/algorithm.
7.4. Visual display of the assessment.
- Corrections in the proposed model (10% bonus) 8.1. What corrections are required in the proposed model to improve its performance? 8.2. How can the model be implemented in the real system of the organization?