Seeking Guidance for Simple yet effective approch for ABSA for my FYP

[watertiger01]

Hi Yang,
I am currently doing ABSA for my Final year Project. Currently my approch is to extract aspects via Rule Based. But it does not work that significantly as compared to other approches. I also tried the word embedings such as word2vec and BOW passing it LDA model for Topic modeling. With that topic being obtained i used sentic algorithm to find the sentiment score of each topic. But the problem was LDA did give me the exact topic words. I kinda feel now that the Deep learning works well for these process but i don't which model to select and where to start.
As you have expertise in this field, i would like to know a simple yet effective appproch to deal with ABSA. If possible can u give me a step by step approch and algorithm to use for ATE (Aspect Term Extraction) aswell as ASA(Aspect Sentiment Analysis).

Thank you