Search for review articles - as examples for comparison - written by other scientist and published in journals of high quality. You might even find reviews on topics similar to yours. Decide whether you are intending to write a "mini-review" or a full-size paper, and collect a substantial number of original publications on the topic you are interested in. Make sure that the most important original papers are included in your reference list. Go back quite a number of years with your selection of relevant papers. Often you will be surprised by your finds. Read the full texts of the original papers, not just the abstract. Often reviews are ending with "Conclusions" and "Outlooks". This could even be of advantage when you apply for research money.
I have got some key points regarding review article, these are given as below:
1. first you should decide which kind of review article you are going to write, weather it is Systematic or general review article.
2. in general review article, you should explore the area which you are going to target in your review article, i mean that you should explore various techniques, approaches, tools and methodology used in that domain by various researchers.
3. discuss each review paper in such a way to extract at least three important points i.e. methodology used, its significance and limitations in that approach.
4. At the end of each part you should provide some concluding remarks and implications in light of explored literature.
5. as far as systematic review is concerned, you should focus on the questions you are gonna search in these articles, and should present the data in both tabular and statistical form. you should follow the rules and methodology of systematic review article where there you need proper inclusion/exclusion criteria, selection of questions and discussion of results being inferred.
Recommender systems have been introduced in various domains, first you will have to decide the field which you are going to explore. It can be movies, research articles, restaurants etc. You need to explore various techniques (CB, CF, Hybrid etc.) various profile matching techniques, dataset being used, evaluation approach being used, and problem addressed etc. You should explore some other aspects as well but this was a kind of example as far as recommender systems are concerned .
@zafar Ali Sb, JazakAllah.. your both the answer is really encouraging and knowledge and idea pouring...Thanks again, hope to get tips and helps from you in future as well.
Brother, you are always welcome, feel free to share knowledge with me. My field of specialization is also recommender systems, I would like to share ideas with you and get help from you as well.