Proposed Algorithm

Factors affecting performance in Group Recommendation

Social Relationship Descriptor
Group Decision Strategies
Expertise Descriptor
Dissimilarity Descriptor
Sparseness in Data

Flowchart

Flow Chart

Factor wise Explanation

Social Relationship Descriptor = IOU of Group Items

Group Decision Strategies: Select Maximum Satisfaction / Average Satisfaction/ Minimum Misery (based on Social Relationship Descriptor)

Expertise Descriptor = Weight assignment based on the highest number of movies watched by a group member

Dissimilarity Descriptor = Dissimilarity based on Average Pairwise Dissimilarity (APD) or Variance Dissimilarity (VD)

Sparseness in Data = Matrix Factorization Experiments

Formulae

Group Decision Strategy: Maximum Satisfaction

Group Decision Strategy: Average Satisfaction

Group Decision Strategy: Minimum Misery

Where,

Social Relationship Descriptor

Expertise Descriptor

Dissimilarity Descriptor:

Average Pairwise Descriptor (APD)

Dissimilarity Descriptor:

Variance Descriptor (VD)

Sparseness Descriptor:

Default MF using the Exponential Decay Learning Rate Scheduler.

Default MF using the Exponential Decay Learning Rate Scheduler with steps.

Baseline-included MF using the Exponential Decay Learning Rate Scheduler.

Baseline-included MF using the Exponential Decay Learning Rate Scheduler. with steps

Default MF refers to:

Error calculation at Gradient Descent=(p_u.q_i - r_u,i)

where

p_u= User Latent Factor

q_i= Item Latent Factor

r_u,i= Training matrix rating for user u and item i

Baseline Included MF refers to:

b_u,i= U + b_u + b_i

Error calculation at Gradient Descent=(b_u,i + p_u.q_i - r_u,i)

where

b_u,i= Baseline Estimate

U= Overall mean rating for all items

b_u= Rating deviation for user u=Average rating of user u - U

b_i= Rating deviation for item i=Average rating of item i - U

p_u= User Latent Factor

q_i= Item Latent Factor

r_u,i= Training matrix rating for user u and item i

Group Concensus Heuristic Function

Valid Combinations Filtering

Explanation of the Flags:

Social_descriptor_enable=False #0

Social_assumption_similar=(Social_descriptor_enable==False) and False #max satisfaction #1

Social_assumption_diverse=(Social_descriptor_enable==False) and False #avg satisfaction #2

Social_assumption_dissimilar=(Social_descriptor_enable==False) and True #min misery #3

Expertise_descriptor_enable=True #4

Dissimilarity_enable=True #5

Dissimilarity_APD_enable=Dissimilarity_enable and False #6

Dissimilarity_VD_enable=Dissimilarity_enable and True #7

Dissimilarity_Jaccard_enable=Dissimilarity_enable and False #8

Dissimilarity_Pearson_enable=Dissimilarity_enable and False #9

Flags=[Social_descriptor_enable (0)

,Social_assumption_similar(1)

,Social_assumption_diverse(2)

,Social_assumption_dissimilar(3)

,Expertise_descriptor_enable(4)

,Dissimilarity_enable(5)

,Dissimilarity_APD_enable(6)

,Dissimilarity_VD_enable(7)

,Dissimilarity_Jaccard_enable(8)

,Dissimilarity_Pearson_enable(9)]

Valid Combinations:

-------------------------------------------------------------------------------------------------------------------------------------------------------

Base cases: Social Relationship + Group Decision Strategies + Expertise + APD

Flags=[1,0,0,0,1,1,1,0,0,0] #estimated social

Flags=[0,1,0,0,1,1,1,0,0,0] #max satisfaction

Flags=[0,0,1,0,1,1,1,0,0,0] #avg satisfaction

Flags=[0,0,0,1,1,1,1,0,0,0] #min misery

-------------------------------------------------------------------------------------------------------------------------------------------------------

No Expertise + No Dissimilarity

Flags=[1,0,0,0,0,0,0,0,0,0]

Flags=[0,1,0,0,0,0,0,0,0,0]

Flags=[0,0,1,0,0,0,0,0,0,0]

Flags=[0,0,0,1,0,0,0,0,0,0]

-------------------------------------------------------------------------------------------------------------------------------------------------------

No Expertise + APD

Flags=[1,0,0,0,0,1,1,0,0,0]

Flags=[0,1,0,0,0,1,1,0,0,0]

Flags=[0,0,1,0,0,1,1,0,0,0]

Flags=[0,0,0,1,0,1,1,0,0,0]

-------------------------------------------------------------------------------------------------------------------------------------------------------

Expertise + VD

Flags=[1,0,0,0,1,1,0,1,0,0]

Flags=[0,1,0,0,1,1,0,1,0,0]

Flags=[0,0,1,0,1,1,0,1,0,0]

Flags=[0,0,0,1,1,1,0,1,0,0]

-------------------------------------------------------------------------------------------------------------------------------------------------------

No Expertise + VD

Flags=[1,0,0,0,0,1,0,1,0,0]

Flags=[0,1,0,0,0,1,0,1,0,0]

Flags=[0,0,1,0,0,1,0,1,0,0]

Flags=[0,0,0,1,0,1,0,1,0,0]

-------------------------------------------------------------------------------------------------------------------------------------------------------

Page updated

Report abuse