Aiming at the deficiency that naive Bayes indoor positioning algorithm ignores correlation between access point（ AP） signal

and ultimately lead to the lack of positioning precision loss

a Wi Fi indoor localization algorithm based on principal component analysis combined with weighted Bayes（ PCA-WBayes） is presented. Wi Fi signal strength of each AP at the reference point in off-line phase is aquired

PCA is used for decorrelation

dimension reduction processing

extract main data characteristics

combine reference point position coordinates

construct location fingerprint database; In the online positioning phase

the signal characteristics of the point to be measured are matched in position fingerprint database

the position of the point to be measured is estimated by Bayes algorithm

get the maximum reference point coordinates of the first w posterior

probability

and assign the weight according to posterior probability

and use the weighted result as position of point to be measured. The experimental results show that the proposed PCA-WBayes algorithm has a smaller positioning error than the KNN

Bayes and other commonly used Wi Fi indoor localization algorithms

and precision of the naive Bayes indoor positioning algorithm is improved by 15. 44 %.