SPECTRA Phones: Systematic Personalized Evaluation for Comprehensive Technological Recommendations and Advancements in Phones

Deciding and purchasing phones becomes crucial in tech-illiterate individuals. Buying phones circulates in a usual manner like acquiring it as soon as it is introduced by the sales officers. Having no further knowledge upon buying might lead to a regretful decision. This study presents application of machine learning techniques such as Cosine Similarity, K-nearest Neighbor, Ball Tree, and Compressed Sparse Row (CSR) Matrix Algorithm in recommending smart phones to consumers based on their personal preferences and requirements. The system undergoes the process of Python programming and GUI implementation to achieve the said inclinations and to have a clear and real-time vision of the purchasing process. The selection process of algorithm revolves only on the abovementioned categories. The models were evaluated using precision, recall, and F1-score evaluation metrics. According to the results presented, Cosine Similarity appears to be the most appropriate recommendation model garnering a score of 71.43 %, 100%, and 83.33% for average of precision, recall, and f1-score respectively.