Robbe is a high-performance chinese word segmentation component built on friso. the accuracy of word segmentation reaches 98.4%, and the speed of word segmentation in simple mode is 3.1 M/sec. the complex mode is as follows: 1.4 M/sec (Ubuntu/2G/dual-core general CPU)
Robbe is a high-performance chinese word segmentation component built on friso. the accuracy of word segmentation reaches 98.4%, and the speed of word segmentation in simple mode is 3.1 M/sec. the complex mode is as follows: 1.4 M/sec (Ubuntu/2G/dual-core general CPU) Friso-1.1 comes: 1. fixed a bug that may cause memory leakage. the interface API is changed to solve the problem that the program needs to be called to automatically release the memory. it is automatically processed in friso_next. for details, refer to the official documentation. 2. added the synonym segmentation function to implement synonym matching. (friso. add_syn = 1 needs to be configured in friso. ini)
3. added the stop word filter function, added a new lex-stopwords.lex dictionary, specially stored the stop word dictionary.
4. the internal code of friso is optimized. only 13 words are needed to load all word libraries. xM memory, and the splitting speed is also improved. simple Mode: 3.1 M/sec, and the complex mode is close to 1.5 M/sec.
5. integrated the jcseg dictionary and added some new words.
6.Php5.3 and later versions are compatible with php5.3 and later versions (compiled by 5.4.16 ). very Net user sends mail to me to compile a vc9 version, robbe-1.1 also provides Linux shared file, vc6 version and vc9 version. 7. added the CHANGES document and the README document for test configuration instructions. Thanks for your attention and feedback. |