Abstract
Recent advances in artificial intelligence architectures have provided templates for automated composition in music generation. However, these automated composition techniques often suffer from a lack of emotional expression and structural coherence. This paper proposes a method for the intelligent generation of music melodies in different styles and modes based on the Fibonacci sequence. Specifically, the pitch frequency of musical notes is quantized using A440Hz, and the direction of melody development is represented by chords processed with the Fibonacci transformation matrix. To meet the predefined requirements for style and mode, the Hidden Markov Model and the Viterbi algorithm are employed to find the optimal solution. Using the example of C major and dominant seventh chords, this paper first generates a state transition matrix by calculating the distances between chords before and after applying the Fibonacci transformation matrix. Then, an observation matrix is constructed based on the probabilities of chords derived from a collection of works in C major by Haydn and Jazz compositions. Finally, the Viterbi algorithm is applied to find the music melody segment that corresponds to the maximum probability. The results demonstrate that the intelligent composition based on the Fibonacci sequence exhibits originality, orderliness, and aesthetic value. The practical application of the algorithm validates its applicability and accuracy, providing an effective method for the advancement of intelligent composition. CCS CONCEPTS • Machine learning • Data analysis and prediction • Probability models

This publication has 3 references indexed in Scilit: