Algorithm is a set of rules to be followed when solving problems.
Algorithm need to be programmed classify and process information
The effeciency & Accuracy of the Algorithm are dependent on how well the algorithm was programmed.
How ML different from Treditional programing ?
Traditional Programming :
Data program→ Computer→ Output
ML:
Data Output→ Computer→ Program
How does ML work ?
Industry use case’s for AI & ML
- Insurence : New Product Development, Insurance Market Analysis
- E- Commerce: Product Recommendations
- Gaming Solutions: Gaming Devices
- Social Media : Facial Recognition Ability, Personalizing your news feed
- Transportation : Route Optimization
- Health Care: Predict Chronic Diseases
- Mobile Technology : Voice to Text Feature
ML Roadmap:
- Objective : Important stage is determing the objectives & constraints used in the modelling, objective should be broken down in smaller logical parts. At this stage performance kpi’s & sucess targets should also be determined for project
- Data : Important to assess the available data sources, the volume of data & it’s composition for Machine learning, large of data required for good results. If data volume is not adequate infrastructure to capture more data has to established.
- Training : Training the model, at this stage,data scientists determine the range of factors that may affect the process being modelled this is often an extensive process to ensure important factors are not missed
- Integration :Model is intigrated into the management system.This intigration is simplified by the fact that the model release only one function. It’s predicts (or) recommends
- Monitoring : Use of ML requires constant monitoring of the model quantity & Stability
Keys to Successful AI/ML Projects
- Clear Business need( What problem is it solving?)
- Define Specific Project/ Product ( How will you know it works ?)
- Build (or) Find Tool/ Model( Build or Buy)
- Find the data
- Test& Adjust( Use Interactions to keep improving it)