如何处理余弦相似度的负值

Question

我根据术语计算了我的文档的tf-idf。然后，我应用LSA来减少术语的维度。 ＆＃39; similarity_dist＆＃39;包含负值（见下表）。如何计算0-1？

范围内的余弦距离

DatabaseMetaData metaData = connection.getMetaData();
ResultSet result = metaData.getProcedureColumns(null,"IFSAPP","DECODE",null);
System.out.println(result.getFetchSize());
result.setFetchSize(100);
System.out.println(result.getFetchSize());

Answer 1

cosine_similarity的范围是-1到1

余弦距离定义为：

// **AppDelegate is a singleton class, its is access any where in view controller.**

    **//crete a variable global in AppDelegate. in**

#import <UIKit/UIKit.h>

@interface AppDelegate : UIResponder <UIApplicationDelegate>
@property (strong, nonatomic) UIWindow *window;

@property (strong, nonatomic) NSString *tokenString;
@property (strong, nonatomic) NSString* edriveridString;

// **you can assign the value in AppDelegate.m**

_tokenString = [[temp objectAtIndex:0] valueForKey:@"token"];
_edriveridString = [[temp objectAtIndex:0] valueForKey:@"edriverId”];

// In View Controller where you want to access this value , write.

AppDelegate *del = [[UIApplication sharedApplication] delegate];
NSLog(@“tokenString -> %@“, del.tokenString);
NSLog(@“ edriveridString-> %@“, del. edriveridString);

因此cosine_distance将在0到2的范围内

请参阅https://en.wikipedia.org/wiki/Cosine_similarity

余弦距离是正空间中常用于补码的术语，即：D_C（A，B）= 1 - S_C（A，B）。

注意：如果 必须在0到1的范围内，您可以使用cosine_distance / 2