recommender systems algorithm

• Matrix Factorization
• Bayesian Personalized Ranking
• Deep Factorization Machines

model

• MF

    class MF(nn.Module):
            def __init__(self, num_factors, num_users, num_items):
                super().__init__()
                self.P = nn.Embedding(num_users, num_factors)
                self.Q = nn.Embedding(num_items, num_factors)
                self.user_bias = nn.Embedding(num_users, 1)
                self.item_bias = nn.Embedding(num_items, 1)
  
            def forward(self, user_id, item_id):
                P_u = self.P(user_id)
                Q_i = self.Q(item_id)
                b_u = self.user_bias(user_id)
                b_i = self.item_bias(item_id)
  
                outputs = (P_u * Q_i).sum(axis=1) + np.squeeze(b_u) + np.squeeze(b_i)
                outputs =  outputs.flatten()
  
                return outputs
  
            def recommend(self, x, y):
                user_id = torch.Tensor(np.array([x])).type(torch.LongTensor).cuda()
                item_id = torch.Tensor(np.array([y])).type(torch.LongTensor).cuda()
                P_u = self.P(user_id)
                Q_i = self.Q(item_id)
                b_u = self.user_bias(user_id)
                b_i = self.item_bias(item_id)
  
                outputs = (P_u * Q_i).sum(axis=1) + np.squeeze(b_u) + np.squeeze(b_i)
                outputs =  outputs.flatten()
  
                return outputs
  

• BPR

    class BPR(nn.Module):
        def __init__(self, user_size, item_size, dim, weight_decay):
            super().__init__()
  
            self.W = nn.Parameter(torch.empty(user_size, dim))
            self.H = nn.Parameter(torch.empty(item_size, dim))
            nn.init.xavier_normal_(self.W.data)
            nn.init.xavier_normal_(self.H.data)
            self.weight_decay = weight_decay
  
        def forward(self, u, i, j):
  
            u = self.W[u, :]
            i = self.H[i, :]
            j = self.H[j, :]
            x_ui = torch.mul(u, i).sum(dim=1)
            x_uj = torch.mul(u, j).sum(dim=1)
            x_uij = x_ui - x_uj
            log_prob = F.logsigmoid(x_uij).sum()
            regularization = self.weight_decay * (u.norm(dim=1).pow(2).sum() + i.norm(dim=1).pow(2).sum() + j.norm(dim=1).pow(2).sum())
            return -log_prob + regularization
  
        def recommend(self, u):
  
            u = self.W[u, :]
            x_ui = torch.mm(u, self.H.t())
            pred = torch.argsort(x_ui, dim=1)
            return pred
  

• DeepFM

    class DeepFM(nn.Module):
        def __init__(self, field_dic, emb_dim, num_factors, mlp_dims, drop_rate=0.1):
            super(DeepFM, self).__init__()
  
            self.ind_embedding = nn.Embedding(field_dic, emb_dim)
            self.car_embedding = nn.Embedding(field_dic, emb_dim)
            self.reg_embedding = nn.Embedding(field_dic, emb_dim)
            self.calc_embedding = nn.Embedding(field_dic, emb_dim)
  
            self.fc = nn.Embedding(field_dic, 1)
            self.linear_layer = nn.Linear(1,1)
  
            input_dim = self.embed_output_dim = num_factors*emb_dim
            self.modules = []
            for dim in mlp_dims:      
                self.modules.append(nn.Linear(input_dim, dim))
                self.modules.append(nn.Sigmoid())
                self.modules.append(nn.Dropout(drop_rate))
                input_dim = dim
            self.modules.append(nn.Linear(dim,1))
            self.mlp = nn.Sequential(*self.modules)
  
            self.classify_layer = nn.Linear(1,2)
  
        def forward(self, ind, car, reg, calc):
            x = torch.cat([ind, car, reg, calc],1).to(device)
  
            embed_ind = self.ind_embedding(ind)
            embed_car = self.car_embedding(car)
            embed_reg = self.reg_embedding(reg)
            embed_calc = self.calc_embedding(calc)
            embed_x = torch.cat([embed_ind, embed_car, embed_reg, embed_calc],1).to(device)
  
            square_of_sum = torch.sum(embed_x, 1) ** 2
            sum_of_square = torch.sum(embed_x ** 2, 1)
  
            inputs = embed_x.view(list(x.size())[0],-1)
  
            x = self.linear_layer(self.fc(x).sum(1)) + 0.5 * (square_of_sum - sum_of_square).sum(1, keepdims=True) + self.mlp(inputs)
  
            x = self.classify_layer(x)
            x = torch.sigmoid(x)
  
            return x
  

data

MovieLens 100K Dataset : https://grouplens.org/datasets/movielens/

Porto Seguro’s Safe Driver Prediction : https://www.kaggle.com/c/porto-seguro-safe-driver-prediction/data?select=train.csv