On Guitar Chord And Pop Lyric Generating Algorithms Based On Complex Network

With the development of information technology,more and more musical works are composed,which means that musical data becomes more and more complex and massive.It is very difficult by traditional methods to process massive musical data.Therefore,complex network theory has been introduced into the field of music analysis.Complex network is an important tool for the study of the complex system.It has complex structures,abundant nodes,varied connections,and so on.Musical data is pretty abstract and generalized.Complex network theory can turn it into digital data,which is more efficient to model networks and process massive data.In this thesis,guitar chord network and lyric network are constructed and analyzed.Chord generating algorithm and lyric generating algorithm are proposed.The main contents of the thesis are as follows:(1)The guitar chord generating algorithm:Firstly,the nodes and edges are defined according to the feature of the guitar and the two-tiered network which contains a verse network and a chorus network is modeled due to the tremendous differences between the verse and chorus.Secondly,it is necessary to analyze the structure of the network.It can be concluded that the chord networks are all small-world networks and scale-free networks.The networks are rich-club and assortative,which means that there are several crucial chords always getting together and accompanying in certain stable and common chord progressions.Thirdly,the generating algorithm of guitar chords based on random walk algorithm is designed.Several common chord processions are regarded as the musical motifs and applied to recomposed the accompaniments.In addition,certain bad chord processions are revised.Finally,the simulation results show that the proposed algorithm can help people to break the stereotyped mindset and strike the spark of innovation.(2)The guitar chord generating algorithm:Firstly,the nodes and edges of lyric network are defined.By the constructed network,the structural properties are analyzed.It can be concluded that the lyric network has a short average shortest path and large clustering coefficient.This network is a small-world and scale-free network.Next,syntax and semantics are very important for pop lyrics.So grammar rules and semantic computing methods for the network are proposed.Then,according to the topological structure of the lyric network,the generating algorithm based on random walk is proposed.Finally,because lyrics have both musical properties and literary properties,it is difficult to get the musical motifs.Therefore,some constrains are made,which can improve the quality of the lyrics.Through a number of experiments,it can be summarized that the lyrics generated by the generating algorithm are as beautiful as the lyrics written by the lyricists.