 Today we are going to deal with the methods used to apply machine learning. At the end of this session, a student will be able to demonstrate the areas where machine learning can be applied. How machine learning can be applied to a specific problem? We know that machine learning orientation differs from the normal software development life cycle. And this encounters creation of your particular model which consists of a number of algorithms and then training these particular algorithms by giving them data which might be around 80% of the available data and then testing it on the template that has been generated by the training samples for the data which is called as testing data which is 20% of the data available. Machine learning is an opportunity to use the data generated by your business to anticipate business change and plan for the future. Machine learning is a sophisticated set of technologies because we use specialized algorithms which take this particular data, train them and then test upon this particular available data. The only valuable thing that we find is that you have to tie technology to the outcomes and see that things are generated in situations of change. It is therefore not static and you have to learn more and more from the data that is available to you and you can prepare for a business change. Keeping all this into mind, we prefer machine learning to give us the best predictions. Now let us think and answer when can we apply machine learning strategy. To answer this question, we first have to define the strategy. The strategy involves first understanding the business problem that you are trying to solve. At what level it is, what should be the answers, how we should present our particular results and what has to be predicted from our particular environment. The status of the existing business plays an important role and we have to definitely know who are the existing customers that are engaged in generating our particular results. Our future holding involves which are the customers we are acquainted with which help us to buy and we are expecting to be with us for the future. Our customers always going to be happy even if there is a change is what we have to analyze. We have to look at the needs of these particular customers and we see that this business needs are never going to be static. They are dependent about knowledge which are regarding the particular customers and what is in the customer's mind is hidden inside a structured, unstructured and a semi-structured data from our particular data set. We use machine learning techniques to be able to uncover these particular patterns previously not visited or which are hidden in our particular data and are missed before. We select the right machine learning algorithms to generate the best prediction and combine it with the appropriate data sources so that we can generate a particular result. Machine learning has proved to be more efficient when we remove the biases from our strategy. Strategic planning and strategic exercises begin by gaining insights into the customer satisfaction. Is giving a reward to the customer or giving a discount to the customer or giving a free gift to the particular customer most advantages is what we have to analyze. Therefore, we have to analyze what others are doing, where our market is heading, what are our competitive threats, what is the impact on our company and we have to anticipate the sudden emergence of new discoveries or new trends that are there in our particular market. Whether we are going to use a supply chain model or we are going to use a model which is based on new trends that is a blockchain or we are going to understand our particular system is traversing in only a sequential or parallel nature. The biases which we have to deal with are the way the company management looks at the data presented and interpreting of the results through their own lens because their own lens will restrict it to only have those customers which they are actually in day-to-day association with. It is easier to be caught unaware because there won't be any analysis of the change and we know that a very prominent business that is successful in nature has the seeds of change existing in it. The factors that we have to analyze are what are the sources of data, what data we had has been changed and therefore who has manipulated this particular data is the data which is available to us reliable. The most important is the context of the particular data which is used for a particular problem because this gives us more than the successful orientation to generate a particular result. There should be a co-relationship between the data elements which helps us to create a particular association which is a relationship between the conditions that have to be applied to our particular data. We have to create a model by using the most appropriate machine learning algorithms based on the business problem being addressed. For example, if it is a supermarket problem then an association rule algorithm would be more preferred but if it is an optimization problem of optimizing a particular rule from the rule which is involved in our particular systems it is better to use a genetic algorithm system. We have to model this data, we have to train the data and then begin to learn from that particular data to generate a particular prediction finally which is our result. The model you design will present an understanding of the data and an ability to predict the outcomes based on that particular data. For example, you might use semantics and to be more accurate synonyms to be used to search a particular page with the help of a machine learning algorithm. Therefore genetic algorithms can be very easily used in this particular environment but you have to understand the pages which are there on a particular system and add a particular algorithm of machine learning which will be related to the semantic aspect and generate a particular result, it might be a decision tree. For more and more accurate planning we use anomalies which are assumed to be errors and as a general user would reject such particular data but these are actual indications of changes that have been taken in our particular environment and predict a better future. More data is added on to the model trained and analyzed and more data that is added on to a particular system using appropriate machine learning algorithms supported by changes will give more accurate and best outcomes for our particular environment. We have to understand machine learning techniques because they best apply to the advanced techniques to manage growth in our particular environment and keep focusing on emerging opportunities which involve changes in our particular environment so that we are now not outdated but we have a machine learning prediction which is focusing on the present situation. Machine learning leverages advanced algorithms and models to continually train data and therefore these changes which are continuously happening in the training environment will give a better result for the testing with additional data to be imposed on our particular system to begin to apply the most appropriate algorithm to a particular problem giving a best prediction therefore using the right algorithms ingesting the most appropriate data and using the best performing models are our three key features for developing a good machine learning involving technical aspects and generating a particular environment. Automation of this process of modeling involves training of the model being done automatically and testing of this particular model to give accurate predictions. As a reference we have used. Thank you.