User Preference and Reviews Analysis with Neural Networks for Travel Recommender Systems

Abstract

These days' people like to explore numerous places all over the world predominantly which are highly recommended. For instance, people who are new to a particular location regularly choose places to visit manually by typing some of them wished by the user working on search engine applications. It becomes difficult whenever to search manually and plan accordingly which might be not accurate. To overcome the above issues, the travel recommendation system applies sentiment analysis comparing user preferences, number of days, number of people and delivers the top recommended places to the user considering the pre-existing reviews of the places which are recommended. While processing it calculates the similarity between user preferences/inclinations and reviews using sentimental attributes (positive and negative) to match the similarity. Travel recommendation system takes sentiment polarity and similarity value as parameters on the whole, also the days to be visited and suggests esteem places.

A profiled side‐channel attack detection using deep learning model with capsule auto‐encoder network

“…The boosting techniques of gradient descent is used to create connections through the statistical models. [22][23][24][25][26] In this work, a new capsule auto-encoder network model is proposed for detecting the side-channel attack. The complex implementation and limited understanding makes capsule network more complex.…”

“…Other regularization factors in XGBoost regulate the shape and structure of trees, strengthening predictions and broadening the applicability of the method. The boosting techniques of gradient descent is used to create connections through the statistical models 22–26 …”

Forecasting Prices in Financial Markets Using Artificial Neural Networks

“…Reference [6] examined the predictive performance of the five of the above linear and non-linear models to forecast the stock market returns of selected indices from developed, emerging, and frontier markets, concluding that there is no single model that can be applied uniformly to all the markets. Nevertheless, reference [7] found that artificial intelligence models such as ANN performed better than traditional models given the ability of this artificial intelligence technique to learn by training any complex input/output mapping that makes its use valuable and attractive from financial time series forecasting to travel recommendations systems [8].…”