Page 139 - Ai_V1.0_Class9
P. 139

• Accuracy: AI models have reached an impressive 98.6% accuracy in detecting Diabetic Retinopathy, on par with
                   specialist eye doctors.
                    • Deployment: Seventy-one vision centers in rural Tamil Nadu, India are utilising this AI technology.
                 Procedure

                 Trained technicians capture high-quality images of patients' eyes using specialised cameras.
                 The AI system analyses these digital images to identify the presence of Diabetic Retinopathy.

                 This approach accelerates the detection process and ensures timely diagnosis.
                 Technicians can operate the AI system without the need for a skilled doctor, making it accessible in rural areas.

                 Benefits
                 The AI-driven solution significantly benefits rural populations by enabling early detection and treatment of Diabetic
                 Retinopathy. This reduces the risk of severe vision loss.


                         AI Project Cycle Mapping Template


                 AI Project Cycle Mapping Template presents how different stages are related to each other and how the functions
                 performed in every phase forms an input for the next phase.

                 The performed task at one stage forms the root for the next stage.
                 AI Project: Customer churn prediction (identifying at-risk customers who are likely to cancel their subscriptions
                 or close/abandon their accounts.)
                    • Problem Scoping
                      ✶ Identify the problem: The telecommunications company wants to reduce customer churn rates.
                      ✶ Define objectives: Develop a predictive model to identify customers at risk of churning.
                    • Data Acquisition
                      ✶ Gather data sources: Collect customer demographics, usage patterns, service history, and churn status data
                      from the company's databases.
                      ✶ Ensure data quality: Clean the data, handle missing values, and remove duplicates.
                    • Data Exploration
                      ✶ Explore the data: Analyse customer demographics, usage patterns, and churn rates through visualisations
                      and statistical summaries.
                      ✶ Preprocess  data: Simplify numerical features, convert categorical variables, and create new metrics like
                      customer tenure.

                    • Modelling
                      ✶ Select techniques: Choose machine learning algorithms suitable for classification tasks, such as logistic
                      regression, decision trees, and random forests.
                      ✶ Train models: Use the prepared data to train multiple models, adjusting hyperparameters and performing
                      cross-validation to optimise performance.
                    • Evaluation
                      ✶ Evaluate models: Assess the performance of each model using metrics like accuracy, precision, recall, and
                      F1-score.
                      ✶ Compare models: Compare the performance of different models to select the best-performing one for
                      deployment.


                                                                         AI Reflection, Project Cycle and Ethics  137
   134   135   136   137   138   139   140   141   142   143   144