An user based and item based movie rating prediction recommender system based on data provided by MovieLens using memory-based Collaborative filtering technique by utilizing Pearson correlation, Euclidean distances, Cosine distances, and K-nearest neighbors algorithms

A Java console application that implements the factionality of the knn algorithm to find the similarity between a new user with only a few non zero ratings of some locations, find the k nearest neighbors through similarity score and then predict the ratings of the new user for the non rated locations.

In this project, a recommendation system has been built and tested for providing a user with recipes to cook based on his/her pantry availability and cooking preferences. A machine learning approach was used to build the system. The dataset which was used to build the recommendation system consisted of user ratings for recipes, cuisines, ingredients and instructions to be followed in the recipe, etc. This dataset was cleaned to ensure better and more accurate predictions. Cleaning also involved obtaining the cuisine and method of cooking involved. Then, a K-Nearest Neighbors algorithm (KNN) was trained on the dataset with target feature as the cuisine and the ingredients as the input features. Then, when a user inputs his/her pantry and cooking method preferred, the trained model is used to predict a cuisine and then using the cuisine, a set of recipes are recommended based on the user preference. The model was tested using K-fold cross validation and found to be about 75% accurate in predicting the cuisine. Dataset link : https://www.kaggle.com/shuyangli94/food-com-recipes-and-user-interactions IDE used : IntelliJ Libraries used : Pandas, Numpy, Sklearn

The problem of predicting a user's star rating for a product, given the user's text review for that product, is called Review Rating Prediction and has lately become a popular problem in machine learning. In this project, we implement an approach which involves a combination of topic modeling and sentiment analysis to achieve this objective by treating Review Rating Prediction as a multi-class classification problem, and building different prediction models by using Latent Dirichlet Allocation as the underlying feature extraction method with three machine learning algorithms, (i) K Nearest Neighbors, (ii) Multinomial Naive Bayes and (iii) Random Forest. We analyze the performance of each of these models to come up with the best model for predicting the ratings from reviews. We use the dataset provided by Yelp for training and testing the models.

Emotion recognition can be arduous for machine learning algorithms, especially when a multitude of test samples are input from various people. A way to combat this could be the use of ensemble learning. Ensemble learning allows for a combination of multiple machine learning algorithms to come to the most accurate conclusion based upon multiple predictions. In this paper, we devise a method of emotion recognition using ensemble learning of multiple machine learning algorithms from: k-nearest neighbors (KNN), multilayer perceptron (MLP), and convolutional neural networks (CNN). A combination of these relatively accurate algorithms can establish a versatile model for emotion recognition that classifies a plethora of input data. Using ensemble learning, we were able to create a generalized and accurate model for emotion recognition. Using the collection of emotional speech recordings, following a template like the RAVDESS speech data set. Our hybrid model using ensemble learning was able to achieve accuracy ratings of up to 84.2% on the given data set.

A Movie Recommendation System and Rating Prediction using collaborative filtering by implementing the K-Nearest Neighbors algorithm

Using a real data set from Netflix, created a program to get hundreds of thousands of reviews and store them in a data structure, then given a user and a movie is able to make a prediction of their rating within 0.8 stars using a k-nearest neighbors algorithm