Question

我的目标是使用library(visNetwork)的潜在Dirichlet分配输出，使用library(text2vec)构建每个主题最多包含100个频率词的网络地图。

在一开始，我曾考虑过使用术语“共生矩阵”（tcm）输出，但是，

我想不出一种适合visNetwork(nodes, edges)上结果的方法。
如果我可以从tcm构建它，我将不知道如何将这些术语追溯到每个主题。
我不知道如何从模型输出中访问单词频率，我所知道的是，使用lda_model$get_top_words(n = 100, lambda = 1)可以根据我选择的lambda来获取每个主题的单词概率，但是我不是完全知道它是否适合代表每个主题中的单词，请原谅我这么辛苦。

这是我从tcm得到的结果：

                drive_element  housing_portion  pump_housing  setting_unit  supplying_electric_power  (...)
drive_element               2                0             0             0                       0.0   
housing_portion             0                1             0             0                       0.0  
pump_housing                0                0             2             0                       0.0 
setting_unit                0                0             0             0                       0.0 
supplying_electric_power    0                0             0             0                       0.2  
(...)

这是我得到的结果

      [,1]                    [,2]                     [,3]                        
 [1,] "electric_motor"        "power_transmission"     "electric_motor"            
 [2,] "permanent_magnet"      "output_shaft"           "combustion_engine"         
 [3,] "stator_core"           "power_source"           "internal_combustion"       
 [4,] "iron_core"             "control_device"         "internal_combustion_engine"
 [5,] "axial_direction"       "input_shaft"            "hybrid_vehicle"            
 [6,] "compression_mechanism" "drive_power"            "control_device"            
 [7,] "rotating_shaft"        "transmission_device"    "motor_generator"           
 [8,] "outer_peripheral"      "rotation_speed"         "engine_electric"           
 [9,] "main_body"             "power_transmitting"     "engine_electric_motor"     
[10,] "electric_machine"      "transmission_mechanism" "braking_force"   
      (...)                   (...)                    (...)
      [,4]                  [,5]                   [,6]                 
 [1,] "power_source"        "drive_shaft"          "power_tool"         
 [2,] "control_circuit"     "storage_device"       "battery_pack"       
 [3,] "switching_element"   "power_storage"        "tool_includes"      
 [4,] "power_supply"        "driving_shaft"        "power_tool_includes"
 [5,] "direct_current"      "power_storage_device" "hand_power"         
 [6,] "current_power"       "high_pressure"        "hand_power_tool"    
 [7,] "light_source"        "swash_plate"          "tool_holder"        
 [8,] "high_voltage"        "shaft_drive"          "tool_body"          
 [9,] "light_emitting"      "figure_figure"        "electric_power_tool"
[10,] "alternating_current" "drive_motor"          "main_body"       
      (...)                   (...)                    (...)   
      [,7]                    [,8]                     [,9]                 
 [1,] "power_source"          "electric_motor"         "handle_assembly"    
 [2,] "electric_power"        "control_unit"           "assembly_includes"  
 [3,] "power_supply"          "control_device"         "longitudinal_axis"  
 [4,] "external_power"        "motor_control"          "jaw_member"         
 [5,] "external_power_source" "rotational_speed"       "surgical_instrument"
 [6,] "secondary_battery"     "electric_power"         "body_portion"       
 [7,] "fuel_cell"             "command_value"          "anvil_assembly"     
 [8,] "storage_battery"       "rotation_speed"         "distal_portion"     
 [9,] "power_supplied"        "electric_motor_control" "jaw_members"        
[10,] "source_power"          "motor_driving"          "includes_handle"    
      (...)                   (...)                    (...)
      [,10]             
 [1,] "ac_power"        
 [2,] "power_supply"    
 [3,] "dc_power"        
 [4,] "power_conversion"
 [5,] "ac_power_supply" 
 [6,] "high_frequency"  
 [7,] "dc_voltage"      
 [8,] "power_converter" 
 [9,] "power_source"    
[10,] "inverter_circuit"
(...)

带有text2vec的网络地图每个主题最多100个频率词

0 个答案: