Sentiment Analysis of Gastronomic Posts from Colour Palettes and Narrative Content

Abstract

This paper presents an approach for analysing gastronomic images and also their related comments published by the Getcookingcanada Instagram account, which belongs to a cooking school. Our approach processes the published images to calculate the moods that the image can generate depending on its colour palette, and also analyses the comments related to each publication for determining the positive or negative sentiment associated to them. A dataset containing all these data has been built, and then the correlations among the data has been developed in order to explain the relation between the mood adjectives and the result of the sentiment analysis of the comments of the food images. 673 food images were analysed; the data analysis was carried out using the Kruskal-Wallis one-way ANOVA test on ranks and Jonckheere-Terpstra’s test. Our results show that there is a significant difference between the different adjectives in terms of sentiment analysis.